<%BANNER%>

Factors Influencing the Purchase of Team Licensed Merchandise

Permanent Link: http://ufdc.ufl.edu/UFE0022590/00001

Material Information

Title: Factors Influencing the Purchase of Team Licensed Merchandise Comparison of High and Low-Involvement Groups
Physical Description: 1 online resource (133 p.)
Language: english
Creator: Lee, Dong
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: attitude, attributes, identification, invariance, involvement, merchandise, satisfaction, values
Health and Human Performance -- Dissertations, Academic -- UF
Genre: Health and Human Performance thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Based on multiple theories and other concepts that influence product consumption, this study proposed and examined a structural model in an attempt to explain consumption of licensed team merchandise. The empirical findings of the current study suggested that consumers? intention to purchase of licensed team merchandise is affected by various factors including personal values, team identification, brand attitude, past expenditure, perceived product attributes, expectancy disconfirmation, and satisfaction. Further, multi-group (MG) measurement invariance test results suggested that the same proposed constructs were being measured equally in each group. MG structural invariance test results suggested that the structural relationships among the tested constructs did not vary due to the difference in the level of personal involvement. It is worth noting that equality was established in the both measurement and structural model, suggesting stability of the proposed model. It is also worth noting that the test of MG invariance of a model is unique because it has not been a common way of model testing within the context of sport, making this study more worthwhile.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Dong Lee.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Zhang, Jianhui.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022590:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022590/00001

Material Information

Title: Factors Influencing the Purchase of Team Licensed Merchandise Comparison of High and Low-Involvement Groups
Physical Description: 1 online resource (133 p.)
Language: english
Creator: Lee, Dong
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: attitude, attributes, identification, invariance, involvement, merchandise, satisfaction, values
Health and Human Performance -- Dissertations, Academic -- UF
Genre: Health and Human Performance thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Based on multiple theories and other concepts that influence product consumption, this study proposed and examined a structural model in an attempt to explain consumption of licensed team merchandise. The empirical findings of the current study suggested that consumers? intention to purchase of licensed team merchandise is affected by various factors including personal values, team identification, brand attitude, past expenditure, perceived product attributes, expectancy disconfirmation, and satisfaction. Further, multi-group (MG) measurement invariance test results suggested that the same proposed constructs were being measured equally in each group. MG structural invariance test results suggested that the structural relationships among the tested constructs did not vary due to the difference in the level of personal involvement. It is worth noting that equality was established in the both measurement and structural model, suggesting stability of the proposed model. It is also worth noting that the test of MG invariance of a model is unique because it has not been a common way of model testing within the context of sport, making this study more worthwhile.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Dong Lee.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Zhang, Jianhui.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022590:00001


This item has the following downloads:


Full Text

PAGE 1

1 FACTORS INFLUENCING THE PURCHASE OF TEAM LICENSED MERCHANDISE: COMPARISON OF HIGHAND LOW-INVOLVEMENT GROUPS By DONGHUN LEE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Donghun Lee

PAGE 3

3 To all who nurtured my intellectual curiosity, academic interests, and sense of scholarship throughout my lifetime, maki ng this milestone possible

PAGE 4

4 ACKNOWLEDGMENTS It is m y pleasure to express my thank to t hose who helped me comple ting this dissertation. I would like to thank my mentors, Drs. Gale n Trail and James Zhang, who have constantly helped me with patience. I woul d also like to thank my committ ee, Dr. Richard Lutz, Dr. Lori Pennington-Gray for valuable insights. Most im portantly, I would like to thank my parents whose unconditional love enabled me to complete this work. None of this would have been possible without their belief in me. My last than k goes to Kyung Hee for her friendship and love.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................10 CHAP TER 1 INTRODUCTION..................................................................................................................11 Growth of Sport Merchandise Sales....................................................................................... 11 Overview of the Research Problem........................................................................................12 Purpose of Study.....................................................................................................................14 Theory Development..............................................................................................................14 Development of Sport Cons um er Behavior Research............................................................ 16 Overview of the Model...........................................................................................................17 Research Questions / Hypotheses........................................................................................... 18 Practical Implications......................................................................................................... ....20 Delimitations of the Study..................................................................................................... .21 Definitions of Terms........................................................................................................... ....21 Overview / Summary............................................................................................................. .23 2 LITERATURE REVIEW.......................................................................................................30 Theoretical Background for the Model................................................................................... 30 Values Theory.................................................................................................................30 Identity Theory................................................................................................................33 Attitude Theory...............................................................................................................35 Satisfaction Theory.......................................................................................................... 38 Perceived Product Attributes and Past Expenditure ........................................................ 41 Effects of Personal Involvement..................................................................................... 43 Description of Sport Specifi c Variables in the Model ............................................................47 Personal Values and Sport Consumption........................................................................ 48 Team Identification and Sport Consumption.................................................................. 49 Attitude and Sport Consumption..................................................................................... 52 Past Product Experience and Sport Consumption...........................................................53 (Dis)confirmation of Expectanci es and Sport Consum ption........................................... 54 Satisfaction and Sport Consumption............................................................................... 54 Perceived Product Attributes and Sport Consumption.................................................... 55 Personal Involvement in rela tion to Sport Consumption ................................................ 56 Summary.................................................................................................................................58

PAGE 6

6 3 METHODOLOGY................................................................................................................. 59 Introduction and Research Design..........................................................................................59 Population and Sample.......................................................................................................... .59 Sampling and Data Collection Procedures.............................................................................60 Instrumentation................................................................................................................ .......61 Validity and Reliability...................................................................................................61 Item Selection and Elimination Justification................................................................... 62 Measurement Scales............................................................................................................. ..62 Personal Values Scale......................................................................................................63 Team Identification.........................................................................................................64 Attitude toward Brand and Attitude toward Product....................................................... 64 Intention to Purchase....................................................................................................... 65 Past Product Expenditure................................................................................................65 Expectancy (Dis)Confirmation........................................................................................ 65 Satisfaction......................................................................................................................66 Perceived Product Attributes...........................................................................................66 Personal Involvement...................................................................................................... 67 Data Analysis..........................................................................................................................67 4 RESULTS...............................................................................................................................69 Pilot Test Results............................................................................................................. .......69 Main Study Results............................................................................................................. ....70 Psychometric Properties of the Scales.................................................................................... 70 Reliabilities (Construct, In ternal Consistency, and Item -To-Total Correlation)............. 70 Discriminant Validity...................................................................................................... 70 Group Classification........................................................................................................71 Baseline Model Establishment............................................................................................... 71 Multi-group Measurement Invarian ce Test (sim ultaneous CFA)........................................... 73 Multi-group Structural Invari ance T est (simultaneous SEM)................................................ 74 5 DISCUSSION.........................................................................................................................91 Focus of the Study..................................................................................................................91 Overview of the Significant Findings.....................................................................................92 Establishment of the Baseline Model.............................................................................. 92 Comparison with the Existing Literature........................................................................ 94 Multi-Group (MG) Invariance Test.................................................................................96 Comparison with the Existing Literature........................................................................ 98 Practical Implications......................................................................................................... ..101 Limitations and Recommendations for Future Study........................................................... 103 Summary and Conclusion.....................................................................................................104

PAGE 7

7 APPENDIX A PSYCHOMETRIC PROPERTIES (PILOT TEST RESULTS)........................................... 106 B CORRELATIONS FOR LATENT FACT ORS ( PILOT TEST RESULTS)........................ 108 C PERSONAL VALUES ITEMS............................................................................................ 110 D TEAM IDENTIFICATION ITEMS..................................................................................... 111 E ATTITUDE TOWARD BRAND/PRODUCT ITEMS ........................................................112 F INTENTION TO PURCHASE ITEMS................................................................................113 G PAST PRODUCT EXPENDITURE ITEMS....................................................................... 114 H EXPECTANCY (DIS)CONFIRMATION ITEMS.............................................................. 115 I SATISFACTION ITEMS.....................................................................................................116 J PERCEIVED PRODUCT ATTRIBUTES ITEMS..............................................................117 K PERSONAL INVOLVEMENT ITEMS...............................................................................118 L PSYCOMETRIC PROPERTIES FROM THE LI TERATRE.............................................. 119 LIST OF REFERENCES.............................................................................................................121 BIOGRAPHICAL SKETCH.......................................................................................................133

PAGE 8

8 LIST OF TABLES Table page 4-1 Means (M), standard deviati ons (SD), beta coefficients ( ), internal consistency ( ), and averag e variance extracted (AVE) for the scales ( N = 736)........................................ 76 4-2 Correlations among latent variable s of the original m odel: step 0 ( N = 736).................... 78 4-3 Correlations among latent variable s of the baseline m odel: step 1 ( N = 736)................... 79 4-4 T-test between HIand LO W -personal involvement groups.............................................80 4-5 Summary of fit statistics for testing mu lti-group invariance of the proposed model ........ 81 4-6 Multivariate lagrange multiplier test by simultaneous CFA (invariance in factor loadings from step 4) .........................................................................................................83 4-7 Multivariate lagrange multiplier test by simultaneous CFA (invariance in covariates from step 4)................................................................................................................... .....84 4-8 Multivariate lagrange multiplier test by simultaneous SEM (invariance in regression coefficients from step 5)....................................................................................................86 4-9 Differences in the strength of the hyp othesized relationships under HIand LOW personal involvement......................................................................................................... 90

PAGE 9

9 LIST OF FIGURES Figure page 2-1 A conceptual framework for sport merchand ise consum ption: A full structural model.... 24 2-2 A structural relationships among c ons tructs explained by Values Theory........................ 25 2-3 A structural relations hips am ong constructs explained by Attitude Theory......................26 2-4 A structural relations hips am ong constructs explained by Identity Theory...................... 27 2-5 A structural relationshi ps am ong constructs explained by Satisfaction Theory................ 28 2-6 An abbreviated version of the full structural m odel (manifest variables not included)..... 29 4-7 Path coefficients from the model te st in step 2. [ref er to Table 4-5] ................................. 87 4-8 Path coefficients from the m odel test in step 3a and 3b.................................................... 88 4-9 Path coefficients from the model test in step 5.................................................................. 89

PAGE 10

10 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FACTORS INFLUENCING THE PURCHASE OF TEAM LICENSED MERCHANDISE: COMPARISON OF HIGHAND LOW-INVOLVEMENT GROUPS By Donghun Lee August 2008 Chair: James Zhang Major: Health and Human Performance Based on multiple theories and other concepts that influence product consumption, this study proposed and examined a structural model in an attempt to explain consumption of licensed team merchandise. The empirical fi ndings of the current study suggested that consumers intention to purchas e of licensed team merchandise is affected by various factors including personal values, team identification, brand attitude, past e xpenditure, perceived product attributes, expectancy disconfirmati on, and satisfaction. Further, multi-group (MG) measurement invariance test re sults suggested that the same proposed constructs were being measured equally in each group. MG structural invariance test results suggested that the structural relationships among the te sted constructs did not vary due to the difference in the level of personal involvement. It is worth noting that equality was established in the both measurement and structural model, suggesting stability of the proposed model. It is also worth noting that the test of MG invariance of a model is unique because it has not been a common way of model testing within the contex t of sport, making this study more worthwhile.

PAGE 11

11 CHAPTER 1 INTRODUCTION This study exam ined how people consume s port merchandise in a predictable way. A theoretical model was proposed and then was tested with empi rical data. The proposed model consisted of multiple constructs pertinent to sport merchandise consumption. This effort was designed to provide relevant ma nagement and marketing implica tions and to offer practical suggestions. The following is a broad overview of the area of interest. Growth of Sport Merchandise Sales Estim ates of the total production and consum ption of sporting goods and services reached up to $560 billion at the end of the 1990s (Howard & Crompton, 2004). In the retail segment, combined sales of licensed merchandise of the f our major professional leagues and universities in the U.S. have doubled from $5.35 billion in 1990 to $10.95 billion in 1999 (Shank, 2002). Direct sales of merchandise at s ports sites are also a very profit able area in sport. For example, Howard and Crompton (2004) estimated that this se gment reached nearly $9 billion at the end of the 1990s. This market trend has continued to grow as sport licensed merchandise sales in retail settings have reached approximately $13 bill ion in 2005 in North America (Brochstein, 2006). Merchandise sales have increased even in minor league levels of sport. The total year-end licensed merchandise sales in minor league baseball were nearly $40 million in 2004 (Broughton, 2005). Not only are sales of merchandise an important source of revenue generation for many sport organizations, but there may also be a synergistic effect with other financial opportunities within sport. For in stance, it is easy to observe sport fans wearing or possessing several items of sport merchandise (e.g., hats, Tshirts, mugs, etc.) while consuming sport (e.g., attending, participating, or watchi ng sporting games or events). Ther efore, it is evident that sales of sport merchandise contribute to the overall consumption of sport.

PAGE 12

12 Overview of the Research Problem W ithin the context of sport, psychological constructs have been often used to explain general consumption activity such as game atte ndance, participation in leisure activity, and media consumption. However, we recognized severa l problems that inaugurated this study. First, a general problem in sport studies was that ther e are few research finding s that explain specific consumption activity such as sport licensed merc handise purchasing. Second, even if there are a few, those existing studies tend to fail to provide quantifiable information that systematically explains what triggered individuals to consum e sport licensed merchandi se. More specifically, those studies tend to focus on narrow topics by accounting for only a few constructs, which often resulted in explaining only a small amount of variance. For example, personal values are a common construct used to explain consumption behavior because they are central to ones t hought processes and are typically enduring. As a result, they are less likely to be affected by other external sources (e.g., specific consumption situations like retail shopping). However, personal values have not explained a large amount of variance in licensed merchandise consumption beha vior (Lee & Trail, 2007). This might have been because values may not be immediate ante cedents of behavior due to their conceptual abstractness. Thus, there may be many additional constructs that mediate, moderate, or in some way influence consumption behavior that have not been investigated. Within the context of sport, external factors such as promotions, fac ilities, and other types of environmental sources have frequently been recognized as influent ial factors of game attendance (Baade & Tiehen, 1990; Greenstein & Marcum, 1981; Hansen & Gauthier, 1989; Zhang, Pease, Hui, & Michaud, 1995; Zhang, Smith, Pease, & Jambor, 1997). In addition, internal factors such as motives and identification have been rec ognized as influential factors for sport consumption (Branscombe & Wann, 1991; Cialdini, Borden, Thorne, Walker, Freeman, &

PAGE 13

13 Sloan, 1976; Laverie & Arnett, 2000; Sutton, Mc Donald, Milne, & Cimperman, 1997; Trail, Fink, & Anderson, 2003; Trail & James, 2001; Wann, 2002; Wann & Branscombe, 1993). However, there may be other variables that in fluence the aforementione d internal factors and determine the impact of extern al factors on behaviors among sport consumers. There has been practically no direct application to sport merchandise consumption that incorporates various psychological constructs such as personal values, identity, at titude, satisfaction, personal involvement, attributes, and intention. Furthermore, when researchers develop a model in an attempt to explain various behaviors, they frequently assume that the model fits sim ilarly for all people. That is, researchers may not take into account that the model as a whole ma y be moderated by some factor that has been known previously to influence relationships within the model. This is such a case. Previous research has shown that level of involvement often moderates relationships among antecedents to consumption. For example, res earchers asserted that due to varying levels of personal involvement, there could be differences in an individuals level of attention and comprehension (Celsi & Olson, 1988), perception (Krugman, 197 7), expectations (O liver & Bearden, 1983), satisfaction (Shaffer, & Sherrell, 1997), etc. In addition, it is possible that the influence of personal involvement on the relationships among f actors such as attitude, satisfaction, identity, etc. could vary, possibly changi ng behavioral intentions or be havior (e.g., product purchase) as well. Celsi and Olson (1988) indicat ed that higher personal releva nce often results in increasing the level of attention and comprehension, increa sing the likelihood of overt behaviors such as shopping and searching. Thus, statis tical examination of a model while incorporating level of personal involvement as a moderator would determin e if the structural paths of the model vary due to personal involvement. It is worth noti ng that testing a model with multiple samples has

PAGE 14

14 rarely been done at a domain leve l, within the context of sport in particular, making the current study more worthwhile (refer to detailed review of involvement in the Literature Review section). From the managerial point of view, desp ite its importance as a source of revenue generation for sport organizations, there has been a lack of sport marketing research pertaining to the consumption of licensed merchandise. Not only does this deficiency of research make this study a worthwhile attempt but ther e also are theoretical as well as empirical aspects that may provide more rationale for this type of research. As a result, this study attempted to develop a comprehensive model that explains a specific consumption activity in a systematic way. Base d on these frameworks, developing and testing a theoretical model will give researchers a chance to f ill the void in the literature that is especially relevant to a contextual level of consumption activity such as the purchase of sport licensed merchandise. Purpose of Study The prim ary objectives of this study were to (1) propose a model that consists of latent constructs that explain consump tion of sport merchandise, (2) stat istically examine the structural relationships among the constructs within the pr oposed model, (3) observe differences in the strength of the relations among the latent constructs between highand low-personal involvement groups, and (4) provide relevant ma nagement and/or marketing implications and offer practical suggestions. The following section discusses theories that support the proposed relationships between each late nt construct that are conceptu ally linked to each other. Theory Development Many conceptualizations have attem pted to elucidate various human behaviors. Those include values theory (Rokeach, 1973a, 1973b), iden tity theory (Stryker, 1968), attitude theory

PAGE 15

15 (Fishbein & Ajzen, 1975), satisfact ion theory (Oliver, 1980), etc. Numerous researchers and scholars have studied personal values as a central theme that influences various behaviors in social science (e.g., Belk, 1984; Braithwaite & Law, 1985; Kahle, 1983; Richins, 1994; Rokeach, Schwartz, 1992, Sweeney & Soutar 2001). Identity was studied to explain behaviors such as role-related behaviors and pr oduct consumption (Burke & Reit zes, 1991; Cialdini et al., 1976; Stryker, 1968, Vinson, Scott, & La mont, 1977). Other researchers a nd scholars have claimed that attitude plays a significant role in brand c hoice and product consumption (Ajzen & Fishbein, 1977; Eagly & Chaiken, 1993; Fish bein & Ajzen; Homer & Ka hle, 1988; Katz, 1960; Lemon, 1973; Lutz, Mackenzie, & Belch, 1983; Mitchell & Olson, 1981). More recently, satisfaction has become a crucial topic that has played a sign ificant role in explai ning various consumption behaviors (Oliver, 1989, 1997a, 1997b; Westbrook & Oliver, 1981). Other premises, including attributes (e.g., Lutz, 1977) and involv ement (Krugman, 1967; Petty & Cacioppo, 1981; Rothschild, 1984), have also re ceived significant attention fro m various disciplines. Although each premise contributes independently to explaining behaviors, many of these concepts are thought to be related to each other. In additi on, regardless of each premises contribution to explaining behaviors, adopting a multifaceted appr oach would be more appropriate due to the fact that human behaviors can be interpreted from many different view s. Empirical findings support the argument that a combination of various factors would predict behaviors better. For instance, Feather (1982) examined attitude and expectancy toward an action, in regard to volunteer actions for movement, and found that a combination of attitudes and expectancies explained more variance (r2 = .452) than separate measures of both attitude and expectancy ( r2 = .380 and r2 = .116), respectively. Although th is research finding may only provide an example of such cases, it is reasonable to conclude that although each construct contributes a conceivable

PAGE 16

16 route to explain consumption behaviors, a single construct (e.g., persona l values) cannot always be an immediate predictor of another (e.g., product consumption). The following section discusses antecedents and consequences of produc t consumption within the context of sport. Development of Sport Cons umer Behavior Research Various theories have been applied to explain different criterion variables (e.g., identification, attitude, satisfaction, gam e attendance, product consumption, and media consumption) within the context of sport. For ex ample, researchers within the context of sport frequently indicated that team identification and behavioral intenti on are two common factors that have disparate effects on sport consum ption (e.g., Madrigal, 2001; Trail & James, 2001; Trail, Anderson, & Fink, 2000, 2005; Wann & Robinson, 2002). Attitude has been studied as an influential factor for sport consumpti on (e.g., Cunningham & Kwon, 2003; Mahony & Howard, 1998; Mahony & Moorman, 1999). More specifically, certain valuerelated behaviors are often influenced by mediating functions. Several resear chers investigated such mediating functions of attitude on various sport beha viors such as televised game watching (e.g., Mahony & Howard, 1998; Mahony & Moorman, 2000), sponsored pr oduct purchases (e.g., Irwin, Lachowetz, Cornwell, & Clark, 2003; Kuzma, Veltri, Kuzm a, & Miller, 2003; Roy & Graeff, 2003) and game attendance (e.g., Cunningham & Kwon, 2003). Numerous researchers (e.g., Laverie & Arnett, 2000; Leeuwen, Quick, & Daniel, 2002; Ma drigal, 1995; Trail, Anderson et al., 2000; Trail, Anderson et al., 2005; Zh ang, Smith, Pease, & Lam, 1998) studied satisfaction to explain game attendance. Behavioral intention has been frequently reported as a consumption behavior in the sport domain (e.g., Cornwell & C oote, 2005; Cunningham & Kwon, 2003; Mahony & Howard, 1998; Mahony & Moorman, 1999; Murray & Howat, 2002). Another example that may explain a specific sport behavior (e.g., sport merchandise purchasing) would be a situation in which product at tributes (such as price of a product, aesthetic

PAGE 17

17 appearance of a product, or qual ity of a product) influence co nsumers choice of one product over another (Lee, Trail, Kwon, & Anderson, 2006). Likewise, there could be many factors that have either direct or indirect (mediated by other factors) influence on actual sport behaviors (sport product consumption). Along with this information, researchers from various areas, including marketing, stress the importance of studying psychologica l constructs such as personal values, attitude, identity, satis faction, and personal involvement in more specific contexts. Accordingly, considering the information discusse d so far, a structural model was proposed, and the following section illustrates the hypot hesized relationships in the model. Overview of the Model This study p roposed a model that is ba sed on values theory (Rokeach, 1973a, 1973b; Schwartz, 1986; Kahle, 1983), iden tity theory (Stryker, 1968), att itude theory (Fishbein & Ajzen, 1975), satisfaction theory (Oliver, 1980), and several other specific concepts shown to influence consumer behavior (i.e., perceive d product attributes). The proposed model consists of two major parts. First, the model explains the latent structural relationships flowing from values to attitudes to behavioral intention (i.e., purchase intention toward produc t and brand). Then, the model explains the influence of satisfa ction (i.e., the disconfirmation or confirmation of expectancies about the purchase and satisfaction with the pu rchase) and perceived product attributes (i.e., perceived benefits of products) on the formation of an attitude. In the former premise, attitude was further cla ssified into two aspects: attitude toward a brand (e.g., Nike) and attitude toward the product (i.e., team-licensed merchandise). The model depicts that the influence of personal values on the formation of attitude toward the product may be mediated by both identification (e.g., identification with a team) and attitude toward the brand. In the latter premise, it was further hypothes ized that an individual may have a perception about product attributes based on satisfaction with prior purcha se. The perception about product

PAGE 18

18 attributes was hypothesized to influence the fo rmation of attitude toward the brand and the product. Bagozzi and Warshaw (1990) argued that situatio nal contingencies such as scarce supply, scarce resources, time constraint s, lack of willpower, and unconscious habits may prevent one from consuming. However, the proposed mode l purposefully excluded the probable impact of contingency factors on behavior because the ultimate consumption behavior this study focuses on is deliberate action. In other words, this study has been developed based on an assumption that a behavior (a purchase) takes place even if situational contingencies may exist to prevent individuals from the action. Research Questions / Hypotheses This study sought to answer the following hypotheses (H) related to three research questions (R1-3). (R1): Do structural relations hips exist among the following constructs [Personal Values (PV), Team Identification (TI), Attitude toward Brand (AB), Attitude toward Product (AP), and Intention to Purchase (IP)]? (R2): Do other factors influen ce the formation of AP [Past expenditure (PE), Expectancy Disconfirmation (ED), Satisfaction (SA), and Product Attributes (PA)]? (R3): Do other factors influence the forma tion of AB [Satisfact ion (SA) and Product attributes (PA)]? In addition, it was hypothesized that due to the moderating functions of the involvement, strength of the relationships am ong the variables would differ for high and low personal involvement subjects (Table 9 illustrates anticipated differences in the strength of the hypothesized relationships with in the proposed model). Thus it was hypothesized that:

PAGE 19

19 R1: H1: PV will directly influence TI. The re lationship between PV and TI will not be influenced by varying degrees of personal involvement. H2: PV will directly influence AB. The re lationship between PV and AB will not be influenced by varying degrees of personal involvement. H3a-b: PV will influence the formation of AP : (a) directly or (b) mediated by AB. The relationship between PV and AB will not be influenced by varying degrees of personal involvement. H4 a-b: TI will influence AB: (a) directl y, or (b) mediated by PA. The relationship between PV and AB will be stronger under low personal involvement. H5: TI will directly influence PA. The relatio nship between PV and AB will be stronger under low personal involvement. H6 a-b: TI will influence AP: (a) directl y, or (b) mediated by AB. The relationship between PV and AB will be stronger under high personal involvement. H7: AB will directly influence AP. The relatio nship between PV and AB will be stronger under low personal involvement. H8: AP will directly influence IP. The relatio nship between PV and AB will be stronger under high personal involvement. R2: H9: PA will directly influence AP. The relatio nship between PV and AB will be stronger under high personal involvement. H10: PE will directly influence PA. The rela tionship between PV a nd AB will be stronger under high personal involvement.

PAGE 20

20 H11: ED will directly influence SA. The rela tionship between PV a nd AB will be stronger under low personal involvement. H12a-b: SA will influence AP: (a) directly or (b) mediated by AB. The relationship between PV and AB will be stronger under low personal involvement. H13: SA will directly influence PA. The rela tionship between PV a nd AB will be stronger under low personal involvement. H14: SA will influence IP: (a) directly, or (b) mediated by AP. The relationship between PV and AB will be stronger under low personal involvement. R3: H15: SA will directly influence AB. The rela tionship between PV a nd AB will be stronger under low personal involvement. Practical Implications Although many of the discussed theories have been applied to explain a form of sport consumption (frequently game attendance), de veloping a model that incorporates various independent constructs within th e licensed sport merchandise cont ext seems crucial to increasing applicability. Comprehensiveness of such a mo del would allow scholars and researchers to develop theories that explain various commonly occurring phe nomena within the domain. In addition, this research effort will enable retail ers to more effectively communicate with sport consumers, which in turn will increase overall sales of sport merchandise. In particular, testing the proposed model under two conditions, high a nd low personal involvement makes this study noteworthy and will help sport merchandise sell ers to better understand their impact on the overall sales. More specificall y, results of this study revealed effects of personal involvement (Petty & Cacioppo, 1981) on the inte ntion to consume a product in that an individual seemed to

PAGE 21

21 process information through different routes (central for high personal involvement and peripheral for low personal involvement), due to the varying levels of consumers involvement. Thus, theoretical model development that in corporates various relevant theories and empirical tests at domain levels seemed to be noteworthy. The findings of this study revealed whether each of the proposed theoretical constructs contributes to elucidate a specific consumption activity, purchasing sport merchandise Given the influence of merchandise sales on the overall sport industry, mark eters of sport merchandise should search for ways to promote personal values, team identification, attitude, sa tisfaction of consumers, and attributes of products. Delimitations of the Study One of the delim itations of the study, as imposed by the researcher, was that there was no attempt to distinguish the difference between s ports fans and spectators as sport merchandise consumers. Also, the exclusion of the effect of attitude toward advertising itself was a further delimitation, as it was beyond the scope of this investigation. Lastly, this study was also delimited by the exclusion of other likely latent influences on th e participants sport merchandise consumption (e.g., situational variables such as product arrangement, in-store information, as well as other interpersonal variable s such as personality, and goals). Definitions of Terms ATTITUDE. a learned predisposition to respond in a consistently favorable or unfavorable manner with respect to a given objec t (Fishbein & Ajzen, 1975, p. 6). ATTITUDE TOWARD BRAND. recipients affective reactions toward the advertised brand or, where desirable, attitude toward purchasing the brand (Lutz, MacK enzie, & Belch, 1983, p. 533).

PAGE 22

22 ATTITUDE TOWARD PRODUCT. In the context of this study, as assessment of perceived product attributes that are available from a ny external sources pertinent to the product. BEHAVIORAL INTENTION. a persons subjective probability that he will perform some behavior (Fishbein & Ajzen, 1975, p. 288). EXPECTANCY DISCONFIRMATION. perceived satisfaction defic it (surplus) after the product experience (Oliver & Linda, 1981, p. 89). IDENTITY. internalized role expectations that provide a guideline for interpreting life experiences (Stryker & Burke, 2000, p. 286). TEAM IDENTIFICATION. an orientation of the self in re gard to other obj ects, including a person or group that results in feelings or sent iments of close attachment (Trail, Anderson, & Fink, 2000, p. 165-166). INTENTION TO PURCHASE. recipients assessments of the likelihood that they will purchase the brand in the future (Lutz, MacKenzie, & Belch, 1983, p. 533). PERSONAL INVOLVEMENT. a state of interest, motivati on, or arousal (Rothschild, 1984, p. 216). PERSONAL VALUES. Established beliefs that result in a specific mode of behavior or endstate of existence [that] is preferred to an opposite mode of behavior or end-state (Rokeach, 1973a, p. 25). PRODUCT ATTRIBUTES. The external product features th at enable products to grant the desired benefits (Gutman, 1982). SATISFACTION. the emotional reaction following a disconfirmation experience which acts on the base attitude level and is cons umption-specific (Oliver, 1981, p. 42).

PAGE 23

23 Overview / Summary Chapter 1 introduced background infor mation re levant to the area of interest, sport merchandise consumption. The overall trends in the growth of sport merchandise sales were discussed, and then the research problem was discussed. The pur pose of the study was described, and a brief overview of theories that have become a basis fo r this study was done, which would help readers to understand the area of interest. Description of the proposed model was provided, and specific hypotheses this study sought to answer were presented. In Chapter 2, the theoretical framework fo r the study was presented including values theory (Rokeach, 1973a, 1973b; among others), identi ty theory (Stryker, 1968), attitude theory (Fishbein & Ajzen, 1975), satisfact ion theory (Oliver, 1980), i nvolvement theory (Krugman, 1967), and other variables that have been shown to influence consumer behavior (i.e., perceived attributes of the product). Description of each theo ry was provided and then its relationship with other theories and variables was discussed. Fu rther discussion of th e relationships among theories is made in an attempt to link each theory that would substantiate conceptual themes of the proposed model. Lastly, a literature review specifically pertinent to the context of sport supporting the proposed relationships among the constructs in the model was provided. The methods of the research were presented in Chapte r 3, including various statistical procedures that were conducted to establish both valid ity and reliability of the scales.

PAGE 24

24 Figure 2-1. A conceptual framework for sport merc handise consumption: A full structural model INTENTION to REPURCHASE Personal VALUES ATTITUDE toward BRAND (e.g., Nike) SATISFACTION with prior purchase(s) ATTITUDE toward PRODUCT Perceived product ATTRIBUTES EXPECTANCY (DIS)CONFIRMATION of prior purchase(s) Past (PRODUCT) EXPENDITURE IDENTIFICATION (e.g., with Dallas Cowboys)

PAGE 25

25 Figure 2-2. A structural re lationships among constructs explained by Values Theory INTENTION to REPURCHASE Personal VALUES ATTITUDE toward BRAND (e.g., Nike) SATISFACTION with prior purchase(s) ATTITUDE toward PRODUCT Perceived product ATTRIBUTES EXPECTANCY (DIS)CONFIRMATION of prior purchase(s) IDENTIFICATION (e.g., with Dallas Cowboys) Past (PRODUCT) EXPENDITURE

PAGE 26

26 Figure 2-3. A structural rela tionships among constructs explained by Attitude Theory INTENTION to REPURCHASE Personal VALUES ATTITUDE toward BRAND (e.g., Nike) SATISFACTION with prior purchase(s) ATTITUDE toward PRODUCT Perceived product ATTRIBUTES EXPECTANCY (DIS)CONFIRMATION of prior purchase(s) IDENTIFICATION (e.g., with Dallas Cowboys) Past (PRODUCT) EXPENDITURE

PAGE 27

27 Figure 2-4. A structural relationships among constructs explained by Identity Theory INTENTION to REPURCHASE Personal VALUES ATTITUDE toward BRAND (e.g., Nike) SATISFACTION with prior purchase(s) ATTITUDE toward PRODUCT Perceived product ATTRIBUTES EXPECTANCY (DIS)CONFIRMATION of prior purchase(s) IDENTIFICATION (e.g., with Dallas Cowboys) Past (PRODUCT) EXPENDITURE

PAGE 28

28 Figure 2-5. A structural rela tionships among constructs expl ained by Satisfaction Theory INTENTION to REPURCHASE Personal VALUES ATTITUDE toward BRAND (e.g., Nike) SATISFACTION with prior purchase(s) ATTITUDE toward PRODUCT Perceived product ATTRIBUTES EXPECTANCY (DIS)CONFIRMATION of prior purchase(s) IDENTIFICATION (e.g., with Dallas Cowboys) Past (PRODUCT) EXPENDITURE

PAGE 29

29 Figure 2-6. An abbreviated versi on of the full structural model (m anifest variables not included). TI AB AP PV PA SA PE ED PV CO PV HE PV PA PV AM IP P IP B IP

PAGE 30

30 CHAPTER 2 LITERATURE REVIEW The goal of Chapter 2 was to provide the th eoretical background for the proposed model. The m odel uses values theory (Rokeach, 1973a, 1973b; refer to Figure 2) attitude theory (Fishbein & Ajzen, 1975; Figure 3), identity theory (Stryker, 1968; Figure 4), satisfaction theory (Oliver, 1980; Figure 5), and involvement theory (Krugman, 1967) to explain the relationships among the following latent constructs variables: personal values, team identification, attitude toward brand, attitude toward product, intention to purchase, perceived product attributes, satisfaction, expectancy disconfir mation, and past expenditure. It was recognized that although one theory might better corroborat e a specific phenomenon related to product consumption, a combination of va rious theories might facilitate even greater understanding about a specific consumption activit y. Indeed, our attempt to link the discussed theories was based on an assumption that behavior (or behavioral intention) is a result of various factors. In other words, many of the discussed factors of each theory were related to the same consequence, attitude formation, behavioral intention, or behavior and thus, they were at least related to some extent. To this end, linkages between theo ries were also discussed. Theoretical Background for the Model Values Theory Definitions and origin of theory. P ersonal values are defined as established beliefs that result in a specific mode of behavior or end-st ate of existence [that] is preferred to an opposite mode of behavior or end-state (Rokeach, 1973a, p. 25). Rokeach (1973a) indicated that individuals within a particular culture tend to display similar personal values. Similar values may be formed because of a function of the similar and shared socialization processes of the

PAGE 31

31 individuals living within a part icular culture. The function of a value system of a person may vary due to dissimilar degrees of importance of the value in the persons life. Descriptions of different aspects of theory. Values theory elaborates relationships among a diverse range of values. For example, Schw artz (1996) and Schwar tz and Bilsky (1990) indicated that values can be distinguished from each other based on the motivational types of goals they convey. More specifically, the bene volence value tends to conflict with the achievement value because the former is more concerned with others welfare whereas the latter is more concerned with the self. At the same time, a complementary relationship may also exist between any two values. For example, the achie vement value may be compatible with the hedonism value because both types of values tend to direct the individual to pursue the persons own end state of exis tence (Schwartz, 1992). Rokeachs Value Scale (1968, 1973b) provides a comprehensive list of personal values that are classified into two domains: terminal a nd instrumental. He specifie d that terminal values are ultimate end-goals of existence, whereas inst rumental values are those beliefs that direct behavior toward broader end st ates (Rokeach & Ball-Rokeach, 1989). The scale contains 18 terminal values such as security, social recogni tion, and freedom and 18 instrumental values such as ambitious, obedient, and self-c ontrolled. The listed values have been tested with large samples in a number of studies across multiple countries Personal values have also been studied by numerous other researchers including Braith waite and Law (1985), Sc hwartz (1992, 1994), and Kahle (1983, 1984), among many othe rs. Schwartz (1992) listed 11 motivational types of values, which was represented by multiple values in hi s study. For example, stimulation (a motivational type) was represented by three values (i.e., an exciting life, a varied life, and daring). These

PAGE 32

32 values were mainly derived from the Rokeachs Value Survey (1973a) and others were from the literature. Values have been studied at the domain le vel as well, for example athletics (Trail & Chelladurai, 2002), organizational culture (H ofstede, 1984; Hofstede, Bond, & Luk, 1993; Hofstede, Neuijen, Ohayv, & Sanders, 1990), ma terialism (Belk, 1988, 1985; Richins, 1994a, 1994b; Richins & Dawson, 1992), general shopping in retail settings (Babin, Darden, & Griffin, 1994; Sweeney & Soutar, 2001), and interpersonal relationships (Herche, 1994). Regardless of these attempts to study values at the domain leve l, which tends to be superior in terms of a validity standpoint, there has been limited research to investigate the effect of values on a specific behavior, sport product consumption. Relationships among variables in the theory. Values have often been identified as influential factors for behaviors (e.g., BallRokeach, Rokeach, & Grube, 1984; Braithwaite & Law, 1985; Rokeach, 1968, 1973b; Rokeach & Ball-Rokeach, 1989; Schwartz, 1992, 1994; Schwartz & Bilsky, 1987, 1990; Schwartz & Huismans, 1995; Schwartz & Sagiv, 1995), personal involvement (Celsi & Olson, 1988; Zaichkowsky, 1985, 1994), attitudes (Homer & Kahle, 1988; Vinson, Scott, & Lamont, 1977), etc. Homer and Kahle examined relationships among values, attitudes, and behaviors in a shoppi ng context and asserted that values influence the formation of attitudes and ultimately imp act consumers behavior. Vinson, Scott, and Lamonts three levels of values classificati on (i.e., global, domain-specific, and product evaluation) are somewhat rela ted to Homer and Kahles valu e-attitude-behavior hierarchy. Homer and Kahle suggested that attitudes play a mediating func tion between values (as guiding principles) and behaviors. Rokeachs (1968) distincti on of attitude from values is consistent with this theme in that attitudes differ from values as attitudes are generally situation-specific.

PAGE 33

33 Although values are related to various types of behaviors such as media preferences, leisure activities, or shopping be haviors (Beatty, Kahle, Homer, & Misra, 1985; Kahle, Beatty, & Homer, 1986), the small amount of variance expl ained in behavior does not make a very convincing case for values to be immediate antecedents of behavi ors (Lee & Trail, 2007). More specifically, Lee and Trail examined the relatio nship between personal values/goals and eight criterion measures that consisted of two cognitiv e measures (i.e., general sport fanship and team identification) and six behavioral measures (i.e., televised spor ts viewing, sport merchandise purchasing, readership of print media, game a ttendance, listenership of radio, and internet consumption). They found that the greatest amount of variance explained by the combined set of values on actual consumption behavior (game at tendance) was only about 10 %. No single value explained more than 4%. A possible reason why personal values may not sufficiently explain subsequent sport consumer behaviors by themselves stems from the idea that other variables may mediate the relationship between values a nd the ultimate consumption activities. Identity Theory Definitions and origin of theory. Identities are defined as internalized role expectations that provide a guideline for interpreting lif e experiences (Stryker & Burke, 2000, p. 286). Identity theory is derived from Meads (1934) symbolic interactionism. Meads symbolic interactionism interprets identity as an interaction between social structure and the self, which in turn affects ones behavior. Str yker and Burke asserted that Mead s symbolic interactionism has played a fundamental role for the study of indi vidual role-related behavior, which provides a basis for contemporary identity theory. It is wort h noting that as opposed to the social identity point of view that emphasizes intergroup discrimination (Tajfe l & Turner, 1979), the identity theory that the current study uses as a theoreti cal framework views identity as a self-centered role expectation that is relative to target objects in a given situation (Stryker & Serpe, 1994).

PAGE 34

34 Descriptions of different aspects of theory. An individual tends to verify an identity by comparing his or her identity standards to the relevant self meanings (Stryker & Burke, 2000). An individual may develop multiple identities based on multiple roles the individual chooses to play (Stryker & Burke), for example, academic role s, athletic/recreational roles, extracurricular roles, and dating roles (Serpe, 1987). Any roles that an individual identified will then define who they are (e.g., father, coach, daughter, fan, etc.). In turn, an identity specifically relevant to a behavior is more likely to be activated when the level of commitment is higher (Stryker & Burke). An identity may be more salient than others depending on circumstances (Stryker & Burke). Therefore, the salience of an identity becomes an important pr edictor of a behavior (Stryker, 1968). Relationships among variables in the theory. Although loyalty is not the primary interest of this study, identification and its relationship with other factors may be better understood when loyalty is understood. For instance, customer loya lty is defined as a deeply held commitment to rebuy or repatronize a preferre d product/service cons istently in the future, thereby causing repetitive same-brand or same brand-set purchasing, despite situational in fluences and marketing efforts having the potential to cause switchi ng behavior (Oliver, 1997b, p. 392). This definition implies that loyal individuals tend to display stro ng positive attitudes, which in turn will likely to lead higher behavioral intenti ons to purchase. Having high leve ls of psychological attachment toward a brand (e.g., identifica tion with a team, like Dallas Cowboys) may demonstrate such status of being loyal toward th at particular brand. The influenc e of identity on actual behaviors has been supported by empirical research findi ngs. For example, Stryker and Serpe (1994) indicated that identity salience e xplained 3 to 8% of the variance in time spent in various rolerelated behaviors (e.g., an extracurricular role or an athletic role). The small amount of variance

PAGE 35

35 explained may indicate that identi ty may not be an immediate prec ursor to behaviors. Literature suggests that identification influences per ception (e.g., Wann & Robins on, 2002) and attitude (Madrigal, 2001; Wann & Branscombe, 1995). Boni nger, Krosnick, and Berent (1995) indicated that identification is an anteced ent of attitude importance. Attitude Theory Definitions and origin of theory. E arly in the 20th century, researchers tended to consider values and attitudes as similar concepts. For inst ance, Allport (1937) referred to values as broad attitudes, which are similar to beliefs emphasizing cognitive as pects of an attitude. More recently, a contrasting view has been prevalent among researchers distinguishing attitudes from values. For instance, Rokeach (1968) asserted that values are distinct from attitudes in that personal values are not limited to specific object s and situations, wherea s attitudes are. In a similar manner, Fishbein and Ajzen (1975) defi ned attitude as a lear ned predisposition to respond in a consistently favorable or unfavorab le manner with respect to a given object. In addition, Eagly and Chaiken (1993) simply defined attitude as an evaluat ive tendency (p. 32). Descriptions of different aspects of theory. Attitudes may be learned and thus can be altered by further learning (Eagly & Chaike n, 1993; Lemon, 1973). Eagly and Chaiken further indicated that an attitude is cognitively learned through direct sources such as information related to a specific brand as well as indirect sources such as information obtained from media. Affective responses resulting from cognitive lear ning portray how attributes are evaluated by consumers. Relative benefits of object entities ar e cognitively evaluated first and then expressed with affective descriptors such as like or di slike and favorable or unfavorable. An evaluation process may actually take place when an individu al encounters classes of stimuli (e.g., product features, which we term attribut es in this study) in which an attitude is formed from the responses to the stimuli (Eagly & Chaiken). An individual may either directly or indirectly come

PAGE 36

36 across classes of stimuli which may result in di fferent attitudinal consequences. As such, an attitude may be derived from past behavioral responses (Ajzen, 1991; Eagly & Chaiken). In these circumstances, past experience supplies sources of information that act as a reference for future evaluation. Attitude measures explai n the nature of attitude in that affective dimensions such as like (dislike) or favorable (unfavorable) tend to indicate signs of attitude (Ajzen & Fishbein, 1977). Three components including cognition, aff ect, and conation are known to be the characteristics of an attitude (Lemon). In partic ular, conative components of an attitude are more relevant to behavioral intention. As a result, an individuals attitude toward an object tends to be expressed through varying ways: as a cognitive response, an affective response, and/or a behavioral response (Eagly & Chaiken). Individuals tend to evaluate cla sses of stimuli that they encounter; as a result of this, an attitude is formed from the responses to th e stimuli (Eagly & Chai ken, 1993). Likewise, an attitude may be formed toward a specific entit y, a so-called attitude obj ect (Eagly & Chaiken). Any entities that encourage forming an attitude can become an attitude object. An attitude may be formed toward tangible entities (e.g., a product or a person) or intangibl e entities (e.g., a brand or a behavior). Attitudes toward a brand and product ar e the two foci of interest for this study. Attitude toward a product tends to be relatively concrete in comparison to attitudes toward a brand. Brand attitude has been defined as the recipients affective reactions toward the advertised brand or, where desira ble, attitude toward purchasing the brand (Lutz, MacKenzie, & Belch, 1983, p. 533). Empirical findings supported th e influence of attitude toward brand on purchase intention in that attitude toward brand explained 31% to 74% of the variance in an intention to purchase a general product (toothpaste; Lutz et al.). Similarly, it is hypothesized that

PAGE 37

37 attitude toward a particular pr oduct may also be formed as a function of perceived benefits derived from consuming the product, based on the attributes specific to the product. Katz (1960) and Lemon (1973) discussed the multiple functions attitudes may serve for individuals: utilitarian adaptive functions (facilitating identificati on with objects or behavior), ego-defensive/externalization functions, valueexpressive functions, or knowledge functions (object appraisal functions). Ea gly and Chaiken (1993) tended to agree that such functions may imply motivational aspects of att itudes. Researchers have shown th at consumers form an attitude toward a specific brand such as Nike, which a ffects their purchase of products under the brand (Lutz et al.; Mitchell & Olson, 1981; Oliver, 1999). Similarly, an individual may form an attitude toward a specific product: for example, merchandi se an individual previously purchased such as a sweatshirt or a hat. In turn, attitude toward the product elicits purchase intentions and perhaps an actual purchase. Relationships among variables in the theory. Homer and Kahles (1988) investigation of the structural relationships among personal values, attitudes, and behavior s indicated conceptual continuity or the idea that personal values in fluence attitude, which in turn affects behavior. Personal values explained up to 33% of the variance in attitude, and attitude explained 31% of the variance in behaviors (Homer & Kahle). Howeve r, when direct influence of personal values on behaviors was measured, personal values e xplained only 2% of the variance in actual behaviors. This was a significant decrease from the amount of variance explained in behavior when mediated by attitude (Homer & Kahle). From this informati on, values theory and attitude theory would link to explain product consump tion behaviors. Fazio, Powell, and Williams (1989) research supported the infl uence of attitude on actual pr oduct consumption (i.e., general products such as candy bars and soft drinks) and established a significant correlation between

PAGE 38

38 attitude and behavior. Ajzen and Fishbein (1977) indicated a con ceptual connection among values (subjective norms), attitude s, and intentions in that inte ntion tends to be a function of attitude and values. Thus, attitudes may not directly influence actua l behaviors; instead, behavioral intentions may be an i mmediate precursor for behaviors. Satisfaction Theory Definitions and origin of theory. T he origin of satisfaction theory is difficult to trace, although Oliver (1981) asserted th at both Engel, Kollat, and Black wells (1968) and Howard and Sheths (1969) studies of consum er behavior in a buying context ma y be the very first sources. Among many other conceptualiza tions, Olivers (1981) descrip tion of satisfaction tends to provide a clear meaning: the emotional react ion following a disconfirmation experience which acts on the base attitude level and is consumption-specific (p. 42). Descriptions of different aspects of theory. Oliver (1981) emphasized two core elements of satisfaction including expecta tion and its confirmation process. In a later study Oliver (1997) described a cycle of satisfaction in the process of consumption. Sp ecifically, previous experience sets a standard (expectation). The confirmation or disconfirmation of the expectancy mediates the level of satisfaction, which in turn affects th e formation of an attitude. Formation of attitude then affects ones intention to consume in the future. The status of ones satisfaction with a product is determined when relative product attributes are compared and appraised in accordance with ones prior experience with the product. Accordingly, researchers have commonly agreed that satisfacti on is a function of expectancy disconfirmati on (Oliver, 1980, 1981, 1997b; Oliver & Linda, 1981). Leeuwen, Quick, and Daniel (2002) supported this prem ise in that disconfirmation of preexisting expectations is directly associ ated with customer satisfacti on. Preexisting expectations set a standard for future evaluation, and post-experi ence evaluation may fall s hort of the preexisting

PAGE 39

39 expectation (i.e., negative disconfirmation), meet the preexisting expectation (i.e., confirmation), or exceed the pre-existing expectation (i.e., posi tive disconfirmation). Likewise, disconfirmation of an expectation may occur when there is a di screpancy in an indivi duals mental comparison between actual experience and anticipated prob ability (Oliver, 1981). Th erefore, satisfaction status is determined as a func tion of the disconfirmation process (e.g., satisfied as a consequence of positive disconfirmation or dissatisfied as a consequence of negative disconfirmation). To this end, expectancy disconfirmation is defined as per ceived satisfaction defi cit (surplus) after the product experience (Oliver & Linda, 1981, p. 89). It is worth noting that unlike other constructs (i.e., personal values, identity, personal involvement, and product attributes) satisfaction is distinctive as it is only measured using a variet y of post-exposure variables (Oliver & Linda). The expectancy disconfirmation in a purchase situation may indicate that the meaning of products may vary depending on ones preexisting expectation because the expectation becomes a threshold for consumers to evaluate the pr oduct attributes (Oliver 1980, 1981). An individual may develop an expectation of a product from e ither direct (e.g., personal past experience) or indirect information (e.g., word of m outh, media, product attributes, etc.). Relationships among variables in the theory. Oliver and Linda (1981) demonstrated hierarchical relationships among elements associ ated with satisfaction in that expectancy disconfirmation was more closely correlated with satisfaction than othe r variables including expectation, preference, and intention. The role of past experience as an antecedent to attitude has often been claimed in the l iterature (Fazio & Zanna, 1978). O liver and Linda indicated that intention is a function of e xpectation, disconfirmation, satis faction, and preference. Prior experience that is directly related to a product itself elicits ones e xpectation (Oliver, 1980).

PAGE 40

40 Oliver (1997b) indicated that loya lty is a function of satisfaction. Thus, loyalty is understood as a consequence of the long-term effect of satisfaction (Oliver, 1997b). Sheth (1973) proposed a conceptual m odel explaining buyer behavior. Among the conceptual themes is expectation. Particular ly, Sheth discussed five factors determining differential expectations, whic h subsequently influence buyer behavior (i.e., background of individuals, information sources, active search, perceptual distor tion, and satisfaction with past purchases). Feather (1992) and Fishbein and Ajze n (1975) indicated that both attitudinal factors and normative factors are antecedents of ones expectation, which furthe r yield an impact on behavior. It is true that in order for expectations to be e ither confirmed or disconfirmed, expectations must preexist. This condition may im ply the need of expectation as an independent construct in the proposed model. However, due to the cognitive nature of expectation as an antecedent of satisfaction, it is impractical to separate the ex pectation from the confirmation process. Accordingly, Oliver (1980) indicated th at satisfaction is the function of an additive combination of expectation and resulting discrepancy perceptions. Oliv er and Lindas (1981) research findings supported these relationships in that disconfirm ation of expectancies explained 21% (for male consumers) and 30% (for female c onsumers) of the variance in satisfaction with general products (i.e., sleeping apparel). In the same study, satisfaction explained a large amount of variance (71% for male and 67% for female) in intention to purchase. Oliver and Linda (1981) argued that product sa tisfaction mediates the relationship between expectancy (dis)confirmation and attitude or intention. This relationship also implies a relationship among them (satisfactio n, attitude, and intention), whic h tends to provide a rationale that supports a conceptual link be tween satisfaction theory and att itude theory. When satisfaction was regressed on preference, 2 to 4% of the va riance was explained; however, when satisfaction

PAGE 41

41 was regressed on intention, 46 to 49% of the variance was explaine d. It is apt to describe both theories as pertinent to each other in that u ltimate consequences of both theories tend to be similar, whether behavioral inten tion or actual behavior (relations hips are depicted in the Figure 5). Accordingly, satisfaction theo ry and attitude theory would link to explain sport product consumption behaviors. Perceived Product Attributes and Past Expenditure The influence of personal values, identificati on, attitude, and satisfa ction on behavior and their interrelationships have been discussed. It is worth noting that alt hough they are fundam ental to explaining behavior, there ma y be other factors that also ha ve an influence on individuals consumption behavior, such as perceived product attributes a nd past expenditure. Perceived product attributes. External factors, such as produ ct attributes, have received attention due to the fact that internal factors, by themselves, may not be sufficient in illuminating consumption behaviors. Vinson, Scott, and Lamonts (1977) three-dimensional value classification (i.e., global values domain-sp ecific values evaluative values) provide a potential explanation of the causal influen ce of product attributes on purchase behavior. Evaluative values are specifically related to produ ct attributes in that ones perceived value of a product tends to be an evaluation of obtainable information such as price, craftsmanship, and aesthetics. (Lee, Trail, Kwon, & Anderson, 2006). Various types of attributes are identified as influential factors on product consumption. Those in clude price (Ferber, 1973; Lee et al.; Sheth, Newman, & Gross, 1991; Oliver, 1999), logo (Broughton, 2005), design (Lee et al.), craftsmanship (Lee et al.), durab ility (Ferber, 1973; Sheth, Newman, & Gross, 1991), reliability (Ferber, 1973; Sheth, Newman, & Gross, 1991), and so on. As depicted in the model, it is hypothesized that an individual t ypically establishes a perception about the products at tributes (e.g., expense relative to the perceived quality of the

PAGE 42

42 product) in relation to the level of satisfac tion with the prior purchase(s), which will subsequently influence the formation of attitude toward the brand or product. In other words, rather than proposing product attributes as a c onstruct having a direct effect on behavior, the current study hypothesizes a mediating function of attitude (i.e., attitude toward brand and product) connecting the perceived product attributes and behavioral intentions. This relationship was evident in Mitchell and Olsons (1981) empirical test of the mediating role of attitude in linking product attributes and be havioral intention. More spec ifically, Mitchell and Olson concluded that beliefs about produc t attributes (i.e., softness, convenience, absorbance, price, and color of facial tissue), along with attitude toward the advertisemen t, had a mediating effect on the relationship between advertisement and behavior al intention. Kardes ( 1988) indicated that consumers make inferences about product attributes which result in significant effects on brand attitude. Graeffs (1997) experiment al results partially supported th is claim in that inferences about product attributes e xplained 23% of the variance in brand attitude. Past product expenditure. Influence of previous experien ce on to other variables has been very common in social sciences. Although it is frequently known that attitude (A) mediates behavior (B) or intention to behavior (IB), A-B or A-IB co rrelations have often been low (Wicker, 1969). Researchers indicated that previous experience influences future behavior or intention to behavior by increasing confiden ce (Bennett & Harrell, 1975) and/or expectancy (Howard & Sheth, 1969). Fazio and Zannas (1978) experiment is one of the few studies examining the influence of types of experience (i .e., direct and indirect ) on attitude. They found stronger attitudes among the consumers who had direct experience (trial) than indirect experiences (verbal descripti ons). The attitude-behavior co rrelation was .54 among the group that had the trial, but it was only .20 in the group who did not try the sample quizzes in the

PAGE 43

43 problem solving experiment. Smith and Swinyard s (1983) experiment provided similar results and pinpointed the concept that attitude-behav ior consistency increased among the subjects who had a behavioral experience (product trial) than those who did not (indirect experience through advertising only). More variance was explained ( R2 = .36) in the group that had the product trial than those who only experienced advertising (R2 = .11). Effects of Personal Involvement Definitions and origin of theory. While Sherif and Cantrils (1947) study is known to be the very first work on involvem ent (Kim, Sco tt, & Crompton, 1997), others (e.g., Celsi & Olson; 1988; Krugman, 1967; Petty & Cacioppo, 1981; Petty, Cacioppo, & Schumann, 1983; Rothschild, 1984; Zaichkowsky, 1985, 1986, & 1994) have expanded the study of involvement. Involvement has primarily been identified as an individual difference fact or. Rothschilds (1984) definition of involvement seems to be the most generic one and contends that involvement is simply a state of interest, motivation, or arousal (p. 216). Zaichkowsky (1985) defined involvement with a product as a persons per ceived relevance of the object based on inherent needs, values, and interests (p. 342). Celsi and Olson suggested that involvement is a perception of self-relevance that is instrumental in ach ieving personal goals a nd values. Although these definitions give slightly different perspectives the core theme underlying these many definitions of involvement mirror Rothschilds concept. Descriptions of different aspects of theory. When personal involvement is studied, three aspects are typically considered : persistence, intensity, and in volvement objects. Intensity is expressed as high or low (Rothsch ild, 1984). Persistence is expresse d as enduring or situational. Types are concerned with what objects one is involved with. Houston and Rithschild (1978) introduced two components of personal involvem ent, enduring involvement (EI) and situational involvement (SI). EI is a l ong-lasting and ongoing concern th at influences a consumers

PAGE 44

44 information processing, whereas SI describes a temporary feeling that is related to a particular situation. Both seem to affect a consumers product consumption with varying degrees of impact. Richins, Bloch, and McQuarrie (1992) indicat ed that importance of a product or purchase situation raises the degree of SI. A more expens ive or riskier product or purchase environment is a typical component that raises the importance of a product or purchase. In other words, when a consumer buys an expensive or riskier product the person perceives gr eater degree of product importance. As such, the entire process influenc es purchase decision. Richins et al. successfully demonstrated that when consumers buy cars a nd expensive clothing, th ey felt greater product importance, and this subsequently increased SI. Researchers have often examined personal invo lvement at two levels (i.e., high and low) rather than treating it as a continuous variable. For example, Rothschild (1984) argued that dichotomy is most preferred (p. 216) because an infinite number of hierarchies is not practical. Researchers asserted that there are multiple objects a person can be involved with including advertisements (Krugman, 1967, 1977), purchase decisions (Clarke & Belk, 1978), products (Howard & Sheth, 1969), etc. Zaichkowsky (1994) emphasized the consistent meaning of personal relevance to define involvement and related it to three categories such as personal (involvement as a personal rele vance of an object to the pers on), physical (involvement as a personal relevance of a physical characteristic of an object to the person), or situational (involvement as a personal relevanc e of a particular situation to the person). When an individual is more involved with a product or situation, the person is more likely to perceive greater personal relevance with the product or situation (Zaichkowsky, 1985). Researchers argued that in a broad perspective, personal involvement is a motivational construct that is composed of both cognitiv e and affective components (Park & Young, 1986;

PAGE 45

45 Zaichkowsky, 1994). Park and Young suggested that personal involvement is composed of a cognitive component and an affective component They described the cognitive component as the personal relevance of a br ands functional performance (uti litarian motive). They described the affective component as the pe rsonal relevance of a brands em otional appeal that expresses a self-image (value-expressive motive). Potential c onsequences of a differenc e in the levels (high vs. low) of personal involvement are cognitive se arch types, processing of obtained information, and decision-making (Rothschild, 1984). Relationships among variables in the theory. Literature suggests that involvement is related to personal values (Zaichkowsky, 1985), brand attitude (Park & Young, 1986), intention (Oliver & Bearden, 1983), expectation (Oliver & Bearden), perception (Howard & Sheth, 1969), satisfaction (Oliver & Bearden; Shaffer & Sher rell, 1997), buyer behavi or (Howard & Sheth), decision making (Clarke & Belk, 1978), and kno wledge (Celsi & Olson, 1988), etc. More specifically, Howard and Sheth i ndicated a probable relationship be tween perception of attributes and involvement with products. For instance, th ey indicated that involvement with a product would increase differences in the perception of product attributes, product importance, and commitment to brand selection. Krugman (1977) asserted that information acquisition or its evaluation, which might subsequently influen ce a persons perception, could vary due to a persons high or low involvement with the product He further indicated that this involvement with a product effect is likely to be greater in relation to advertising when the level is high. Due to higher levels of perceived relevance or im portance of a product to the person (which is demonstrated via a greater degree of involvement), involvement theory assumes that the person would thus be more interested in or comm itted to external source of information (i.e., advertising) about the related product.

PAGE 46

46 Celsi and Olsons (1988) desc ription of involvement infe rs a plausible relationship between involvement and other constructs. For ex ample, involvement and its relationship with overt behavior can be inferred from their asse rtion that higher persona l relevance between the self and the product tends to enhance feelings of personal relevance or involvement with the product. In the end, higher person al relevance is more likely to bring added effects on cognitive behaviors such as attention and comprehension processes, which increases the likelihood of overt behaviors such as shopping and s earching (Celsi & Olson). Oliver and Bearden studied the role of involvement in satisfaction processes and c oncluded that involvement is likely to raise expectations prior to product use and have continuous effects in post-usage evaluation. Therefore, it is reasonable to assert that level of satisfaction is at least partially determined by ones level of involvement. Oliver and Bearden indicated that, in gene ral, high involvement groups rated considerably higher on preand pos t-usage variables (i.e., attitude, intention, and disconfirmation of expectation). Personal involvement also influenced the relationship between expectancy disconfirmation and satisfaction among the patients in a large medial clinic. As a result, in high-involvement situations, expectan cy disconfirmation explai ned 10 to 26% of the variance in satisfaction, whereas in low involvement situations, no significant relationship was found (Shaffer, & Sherrell, 1997). Also, Celsi and Olsons empirical findings established that personal involvement accounted for up to 26 % of the variance in cognitiv e behaviors (i.e., the time spent processing the advertisements, and the amount of attention given to the process of product consumption). Most of the social-psychologica l constructs within the proposed model tend to be formed through cognitively or affectively learning (Eag ly & Chaiken, 1993; Fishbein & Ajzen, 1975; Lemon, 1973). Thus, it is reasonable to assume that depending on how an individual is

PAGE 47

47 personally involved with a product, the strength of the relations between one factor and another would vary (refer to Research questions/Hypotheses section for detailed relationships on the page 11-12). As a consequence, this study exam ined the structural relationships among the constructs in the model under two conditions: hi gh (HI) versus low (LOW) personal involvement (PI). The following sections discuss how each construct is particularly related to the domain of sport. Description of Sport Specific Variables in the Model Variables within the proposed m odel have spec ifically been studied by researchers in the context of sport. For example, Kahle and othe r researchers (e.g., Homer & Kahle, 1988; Kahle, Kambara, & Rose, 1996; Kahle, Beatty, & Home r, 1986; Kahle, Duncan, Dalakas, & Aiken, 2001; Lee et al., 2006; Trail & Chelladurai, 2002; Tr ail, Fink et al., 2003) a pplied the concept of personal values in an attempt to explain sport behaviors (e.g., game attendance). Wann and other researchers (e.g., Branscombe & Wann, 1991; Fi nk, Trail, & Anderson, 2002; Funk & James, 2001; Kwon, Trail, & Anderson, 2005; Mahony, Ma drigal, & Howard, 2000; Sutton, McDonald, Milne, & Cimperman, 1997; Trail, Anderson et al., 2005; Wann & Branscombe, 1993; Wann & Dolan, 1994a; Wann, 2002; Wann, Ensor, & Bilyeu, 2001) used social identity theory and/or identity theory as a basic frame to develop team identification and explained various sport behaviors (e.g., game attendance). Res earchers (e.g., Cunningham & Kwon, 2003; Mahony & Howard, 1998; Mahony & Moorman, 1999) have chosen attitude as a variable explaining sport behaviors. Numerous researchers (e.g., Caro & Garcia, 2007; Lapidus & Schibrowsky, 1996; Laverie & Arnett, 2000; Leeuwen, Quick, & Daniel, 2002; Madrigal, 1995; Trail, Anderson et al., 2000; Trail, Anderson et al., 2005; Zhang, Sm ith, Pease, & Lam, 1998) applied satisfaction theory to explain sport behaviors (e.g., game attendance). Behavioral intention has been frequently studied by many researchers in the co ntext of sport as well (e.g., Beatty & Kahle,

PAGE 48

48 1988; Cornwell & Coote, 2005; Cunningha m & Kwon, 2003; Mahony & Howard, 1998; Mahony & Moorman, 1999; Murray & Howat, 2002) Although not as co mmon, other variables including involvement (e.g., Funk, Ridinger, & Moorman, 2004; Iwasaki & Havitz, 1998), attributes (e.g., Lee & Trail, 2007; Madrigal, 2003), and knowledge (e.g. Trail & James, 2001) have also been studied in re lation to sport consumption. Personal Values and Sport Consumption Num erous researchers have stud ied personal values in a gene ral context as noted above. However, the study of personal values and their a pplication in the area of sport is limited. A few exceptions have applied the concept of personal valu es to the domain of sport in an attempt to explain specific behaviors. Those include Ka hle (e.g., Kahle, Duncan, Dalakas, & Aiken, 2001; Kahle, Kambara, & Rose, 1996) and several othe rs (e.g., Lee et al., 2006; Trail & Chelladurai, 2002). Beatty, Kahle, Homer, and Misra (1985) f ound that individuals th at regard hedonistic values higher (fun and enjoyment) are more likely to engage in leisure activities (e.g., skiing and camping). Regardless of the efforts, topics st udied were often narrowly focused, such as on leisure activity, failing to extend its application to other areas su ch as merchandise consumption. The one exception is the research by Lee and Trai l (2007) in that they examined the relationship between personal values and eight criterion measures including two cognitive measures (i.e., general sport fanship and team id entification) and six behavioral measures (i.e., televised sports viewing, sport merchandise purchasing, readership of print media, game attendance, listenership of radio, and internet consumption). They found that correlations between ambition and patriotism values and most of the cognitions and behaviors were significant. Items for ambition values were derived from Schwar tzs (1992) achievement value. It ems for patriotism values were derived from Schwartzs security value. Ambition and patriotism values explained some variance in the following cognitions and behaviors: 9 and 14%, respectively (general sport fanship), 5 and

PAGE 49

49 16% (team identification), 2 and 3% (televised sports viewing), 2 and 2% (sport merchandise purchasing), 5 and 4% (game attendance). Othe r personal values such as conservatism and hedonism were also significantly correlated with at least three or more of the eight criterion measures (variance explained ranged from 2 to 9% and 1 to 3%, respectively). Items for conservatism values were derived from Schwartzs tradition value. Based on these findings, and the research a nd theory presented in the section about personal values that was not sport specific, th e present study hypothesized that personal values would influence the level of identification with a team directly. Furthermore, personal values would influence the formation of attitudes toward brand either directly or mediated by team identification. Personal values would influence th e formation of attitude toward product either directly or mediated by at titude toward brand. Team Identification and Sport Consumption Identity th eory has been used to explain w hy people become a fan of a specific sport team (Trail, Anderson et al., 2005). Trail, Anderson et al. (2000) defined identification as an orientation of the self in regard to other obj ects, including a person or group that results in feelings or sentiments of cl ose attachment (pp. 165-166). From this basis, Trail and James (2001) developed the items for the Team Identifi cation Index (TII) as a cr iterion validity measure for their motivations scale; however, the items were first reported in Tra il, Fink et al. (2003). The TII was created to measure cognitive connection toward a specific sport team. During the last two decades, the impact of identification on various sport consumer behaviors has been a key topic among scholars and researchers in the field of sport. Researchers frequently agreed that team identification pl ays a significant role in enhancing consumers experience of sport (Branscombe & Wann, 1991; Ci aldini et al., 1976; Laverie & Arnett, 2000; Sutton, McDonald, Milne, & Cimperman, 1997; Trail, Fink et al., 2003; Trail & James, 2001;

PAGE 50

50 Wann, 2002; Wann & Branscombe, 1993, 1995; Wann & Dolan, 1994a). Wann and Dolan (1994b) found that participants w ith high identification tended to evaluate a teams performance more favorably (attitude) than did participants with low iden tification, indicating that team identification influences attitude toward a pr oduct, in this case the game itself. Wann and Branscombe (1995) demonstrated structural re lationships between team identification and knowledge about the sports (i.e., basketball), pl ayers, and team history. Team identification explained approximately 23% of the variance in spectators knowledge (Wann & Branscombe). They also found that highly identified fans displa yed positive attitudes (fa voritism) toward other fans that supported the same team. Madrigal ( 2001) argued that the rela tionship between team identification and purchase intention is mediated by attitude toward the purc hase behavior. Trail, Robinson, Dick, and Gillentine (2003) found that casual and devoted fans displayed different levels of team identification and thus engage d in diverse subsequent sport behaviors (the magnitude of commitment in terms of money and time were greater for fans with higher team identification than casual fans). Sutton, McDonald, Milne, and Cimperman (1997) classified sport fans according to the level of team identifi cation (i.e., social fans, focused fans, and vested fans, as the latter displayed stronger team identif ication). They argued that a high level of fan identity would impact sport fans behavior, such as decreasing price sensitivity and performanceoutcome sensitivity. However, this identity-behav ior relationship was not empirically tested in their study. Buying a particular piec e of sport merchandise over a li st of other options may thus reflect the fans identity with the team because sport fans often purchase sport licensed merchandise to support their favorite teams or play ers for themselves or for someone else. Thus, those who display higher levels of identification with a team are more likely to attend games, buy sport merchandise, and support the team (Tra il, Anderson et al., 2005). The influence of

PAGE 51

51 identity on actual behavior was partially support ed by Laverie and Arnett s (2000) research that identity explained 15% of the variance in womens basketball game attendance. Smith, Patterson, Williams, and Hogg (1981) and Sutton, Mc Donald, and Milne suggested that fans with high team identification are more likely to attend games. Kwon et al. (2005) indicated that identity (termed as points of attachment) explai ned 38% of the variance in behavioral intention (conative loyalty) and 25% of the variance in ac tual behavior (game attendance). More recently, Kwon, Trail, and James (2007) tested three models (a direct effect, a pa rtially mediated, and a fully mediated) depicting relationships among t eam identification, perceived value of product attributes, and purchase intention. They repor ted the fully mediated model as the most parsimonious one--indicating that identification with a team influenced intention to purchase licensed-sport apparel, which was mediated by co nsumers perceived valu e of product attributes. Perceived value of product attr ibutes explained a large amount of variance (42.6%) in purchase intention (Kwon, Trail et al.). The influence of team identification on the formation of brand attitude within the context of sport is con ceptually supported by Gladden and colleagues study of brand equity. For instance, based on Keller s (1993) framework of brand association, Gladden and colleagues (Gladden, Irwin, & Sutton, 2001; Gladden & Milne, 1999; Gladden, Milne, & Sutton, 1998) argued that consumers perceived favor ability (note that this is a common measure of attitude) toward a brand constitutes brand asso ciation. In addition, the brand associations are derived from the emotional iden tification with a particular t eam (Gladden, Milne et al., p. 3). Based on these findings, and the research presente d in the section about identity that was not sport specific, the present st udy hypothesized that team identification would influence the formation of attitude toward bra nd either directly or mediated by perceived product attributes. It

PAGE 52

52 was further hypothesized that team identification would directly influence attitude toward the product. Attitude and Sport Consumption The proposed m odel partially explained the latent relationships between attitudes and behavior. In this regard, attitude was further classified into two aspects: attitude toward the merchandise brand (e.g., Nike) and attitude to ward the product (e.g., merchandise individual consumers previously purchased or would purchase in the future such as a sweatshirt, a hat, or a mug). Attitude and its relationship with other variables have often been studied in the context of sport. For instance, Zillman, Bryant, and Sapl osky (1989) indicated that sport fans positive attitudes toward a team resulted in increasing po sitive perceptions of games, which subsequently enhanced fans televised sports viewing choices. Similarly, Mahony and Moorman (2000) indicated that when individuals displayed strong positive attitude toward a team, they were more likely to watch their favorite team on television than the best performer in that league ( p < .01). Fans were also more likely to watch their favorite team on television if they had a positive attitude toward the team (p < .001; Mahony & Howard, 1998). The re lationship of attitude with behavioral intention in the context of sport is also supported by Madrigals (2001) study in that a hierarchical relationship existed among beliefs, attitude, and inte ntion when examining televised sports viewing behaviors. In Cunningham a nd Kwons (2003) study, attitudes accounted for 8% of the variance in intentions to attend a hockey game. Roy and Gr aeff (2003) indicated that sport teams increase their involvement in local charities, based on the premise that a positive attitude toward the team increases th e purchase intention of tickets or logo merchandise. Irwin, Lachowetz, Cornwell, and Clark (2003) provi ded supporting evidence fo r the influence of attitude on sport consumption in that attitude positively infl uenced the purchase of sport

PAGE 53

53 sponsored products ( M = 4.43 out of 5.0). Attitude explained 31 % of the variance in the purchase intentions toward the products of corporate sponsors (Madrigal, 2001). Based on these findings, and the research results discussed in the section about attitude that was not sport specific, the present study hypothe sized that attitude toward brand (e.g., NIKE) would influence the formation of attitude toward product (e.g., jackets, hats, or jersey). It was further hypothesized that attitude toward product would subsequently influence intention to purchase team licensed merchandise in the future. Past Product Experience and Sport Consumption Consistent with Ajzens (1991) them e, it is evident that past behavior serves as an antecedent for attitudes in the context of sport. For instance, Trail, Anderson, and Lee (2006) showed that past experience e xplained approximately 21% of th e variance in preseason team identification and 25% of the variance in behavi oral intentions (future game attendance), but only 2% of actual attendance. Cunningham and Kwon (2003) demonstrated a significant relationship between past experi ence (i.e., game attendance) and behavioral intentions, but the small amount of variance explained (2%) is a concern. The small amount of variance explained may imply that there may be other or mediating factors that better explain consumption activity. The current study, therefore, hypothe sized that past product expenditu re would have an influence on an individuals intention toward a behavior mediated by the individuals attitude toward the product. Various types of previous experiences may bring a differing impact on satisfaction, attitude, or behavioral intention. However, this study purposefully limited the scope of previous experience to past product expenditure to ex plain merchandise consumption activity in the domain of sport.

PAGE 54

54 (Dis)confirmation of Expectancies and Sport Consumption As reviewed earlier, expectancy (dis)confirmati on (ED) is one of the two core elements of satisfaction (SA). In other words, in the process of product consumption, preexisting expectation sets a standard for future evaluation, and then the confirmation of the expectancy mediates the level of satisfaction. Th is ED-SA relationship is often suppor ted in the context of sport. For example, Caro and Garcias (2006) findings supp orted this relationship in that expectancy disconfirmation explained a little over 11 perc ent of the variance in satisfaction (running competition). Madrigal (1995) indicated that expe ctancy disconfirmation on the quality of the game influenced sport fans satisfaction, but the relation was mediated by both basking in reflected glory (BIRG) and enjoyment. Leeuwen, Quick, and Daniels (2002) Sport Spectator Satisfaction Model depicted that disconfirmation of preexisting expectations is an important influence on spectators satisfac tion. However, this relationship was not tested with empirical data. Trail, Fink et al., 2003 determined that di sconfirmation of expectancies explained 32% of the variance in affective state, which was measur ed by satisfaction and positive/negative affect Trail, Anderson et al. (2005) tested three comp eting models that examined relationships among factors including expectancy disconfirmation and satisfaction. In their study, expectancy disconfirmation explained 41% of the variance in satisfaction. Based on these findings, and the research results presented in the section about expectancy disconfirmation that was not sport specific, the present study hypothesized that expectancy disconfirmation would directly influence satisfaction. Satisfaction and Sport Consumption Satisfaction has often been studied to explain sport specta tor behavior. For instance, Madrigal (1995) suggested a cognition-affect-sat isfaction sequence to explain individuals consum ption behavior. Caro and Garcia ( 2007) studied relationshi ps among cognition

PAGE 55

55 (expectancy disconfirmation), emotion (arousal and pleasure), satisfaction, and loyalty of participants in a running competition. Madrigal (2003) proposed a model that depi cted performance satisfaction influencing participants optimism about future performances of the team that they supported. He determined that satisfaction with the teams performance explained 23% of the variance in entertainment values and 15% of the variance in optimism a bout how well the team would do in the future. This finding may suggest that satisfaction will in fluence attitudes positively because affect (i.e., feelings/confidence), which is the main compone nt of attitude measur ement, was used to represent the optimism. Matsuoka Chelladurai, and Harada ( 2003) showed that satisfaction influenced intention to attend future games. More specifically, satisfaction with performance explained 26%, satisfaction with win/loss expl ained 22%, and satisfaction with game quality explained 25%, of intentions to attend future games. Trail, Anderson et al. (2005) found similar results in that satisfaction explai ned 18% of the variance in cona tive loyalty. Regardless of these efforts to relate satisfaction and sport consump tion, many available satisfac tion studies within the context of sport have narrowly focused on game attendance (Cunningham & Kwon, 2003; Funk & Pastore, 2000; Lapidus & Schibrowsky, 1996; Laverie & Arnett, 2000; Leeuwen, Quick, & Daniel, 2002; Madrigal, 1995; Zhang, Smith, Pease, & Lam, 1998) or leisur e activity (Wakefield & Blodgett, 1994). Based on the review of litera ture, the present study hypothesized that satisfaction would have an effect on the formation of attitudes toward a brand or product. In addition, it was hypothesized that co nsumers post-usage evaluation (s atisfaction/dissatisfaction) with a product would influence th e perception about the product. Perceived Product Attributes and Sport Consumption Gut man (1982) referred to product attributes as the product features that enable products to endow the desired benefits to consumers; fo r example, product featur es such as price,

PAGE 56

56 craftsmanship, and aesthetics (B roughton, 2005). Unfortunately, ther e have been very few, if any, studies that have examined attributes a nd their impact on product consumption within the domain of sport. Lee et al.s (2006) study may be one of th e few studies providing information about product attributes (perceived value of product) and their impact on sport product consumption. Lee et al. defined the perceived value of a product as an objects relative worth to the individual. They concluded that unique f eatures of a product elicit consum ption behaviors such as the purchase of licensed sport merchandi se. Product attributes that were important for individuals to purchase a piece of licensed sport merchandise incl uded reasonable prices or quality, nostalgia (personal history), craftsmanshi p, and aesthetic beauty of the product they purchased. Personal history represents a personal e xperience through the products that induce memories of particular sporting events, venues, or pers ons (Lee et al.). More specif ically, products an individual collected may help to recall a particular sporti ng event like a championship game. Craftsmanship and Price/Quality may indicate util itarian aspects of the products in that product quality is an important feature for an individu al to judge product worth rega rding its relative benefits the product offer (Lee et al.). In their study, mean valu es were close to or gr eater than the midpoint of the scale (4.0) for all of th e perceived attributes variable s (Craftsmanship = 5.81, Price = 4.99, Nostalgia = 3.97), except Prestige /Status (2.85). Based on these findings, and the research and theory presented in the section about perceived pr oduct attributes that was not sport specific, the present study hypothesized that sport consumers perception of product attributes would influence attitude toward the product either di rectly or mediated by brand attitude. Personal Involvement in relatio n to Sport Consumption Researchers argue that involvem ent is an im portant individual diffe rence variable that affects various recreation, tourism, and sport activities (e.g., Ball & Tasaki, 1992; Dimanche,

PAGE 57

57 Havitz, & Howard, 1995; Green & Chalip, 1997 ; Iwasaki & Havitz, 1998; Kim, Scott, & Crompton, 1997; Laverie & Arnett, 2000; McInty re, 1989; Schuett, 1993; Keistetter & Kovich, 1997). Laverie and Arnett (2000) argued that high involvement leads to high identity salience, rather than vice versa, by examining structur al relationships between involvement as an important antecedent for other constructs in cluding fan identity, satisfaction, and sport consumption (NCAA womens basketball game attendance). In their study, situational and enduring involvement accounted 8% of the variance in identity salience. Situational involvement in particular explained a large amount of varian ce (26%) in satisfaction. It is worth noting that Ball and Tasaki (1992) provided an important clue to differentia te attachment, which is often understood as a form of identification, from invol vement by pointing out th at involvement deals with a product category, rather than a posse ssion (p. 159). Relatedly, Laverie and Arnett illustrated watching a basketball game as a product category and a particular team as a specific product. Green and Chalip (1997), however, provided a contrary result through their theory that involvement (i.e., childrens enduring invol vement in youth soccer) was influenced by satisfaction (R2 = 51%) rather than involvement influencing satisfaction. According to their findings, ongoing interest (enduring involvement) in youth soccer is established based on the stakeholders satisfaction with their participation in the activity. Regardless of these research efforts, empiri cal examination of personal involvement in relation to sport consumption has been very lim ited. Among the available lit erature, the majority of the involvement studies have narrowly focuse d on explaining leisure activities, thus, limiting its application to other areas. This limitation necessitates further res earch in linking personal involvement in relation to a specific domain-re levant activity such as sport merchandise

PAGE 58

58 purchase. In addition, within th e domain of tourism, recreation, and sport, there has been no study that investigates structural relationships among various f actors and examines involvement effect under different conditions simultaneousl y (high and low personal involvement). This scarcity further makes this study noteworthy. Summary In summ ary, based on values theory, identity theory, attitude theory, satisfaction theory, and other variables that influence consumer be havior, this study propos ed a conceptual model that attempted to explain sport merchandise consumption behavior. The proposed model was empirically tested to provide evidence to suppor t the structural relationships among the proposed constructs in the model. The propos ed model was depicted as follows: Attitude toward a product (AP) is a function of past expenditure (PE), expectancy disconfirmation (ED), satisfaction (SA), and percei ved product attributes (PA). Attitude toward brand (AB) is a function of per ceived product attributes (PA) and satisfaction (SA). Intention to purchase (IP) is thus a function of personal values (PV), team iden tification (ID), attitude toward brand (AB), attitude toward pr oduct (AP), and satisfact ion (SA). The following is an abbreviated version of the full structural model.

PAGE 59

59 CHAPTER 3 METHODOLOGY Introduction and Research Design This study had three objectives: (1) to propose a model that consists of latent constructs that explain consum ption of sport merchandise, (2) to statistically examine the structural relationships among the latent constructs within the model, and (3 ) to observe differences in the strength of the relations among the latent constructs between highand low-personal involvement groups. In order to fulfill these objectives, the methods sec tion consisted of two phases: a preliminary study and a main study. To accomplish the objectives, several phase s were conducted including item generation, item validation (pilot test), and data analysis. Items for each construct [personal values, team identification, attitude toward th e brand, attitude toward the pr oduct, purchase intentions, past product expenditure, expectancy (dis)confirmation, satisfaction, perceived product attributes, and personal involvement] were identified from the literature. A panel of experts was asked to check content validity of the items relevant to the constructs. Confirmatory factor analysis was conducted to measure reliability (construct reliability) and valid ity (discriminant validity). Finally, using Equations (EQS; Bentler, 2003), a structural equation modeling (SEM) analysis was conducted to examine structural relationships among all the constructs. In addition, separate SEMs were conducted for the highand lowpersonal involvement groups. The following provided more specific informa tion regarding each procedure. Population and Sample Student sam ples were used in both phases of the study. A student population is a viable sample frame for this type of study because st udents are often a major segment of the population of sport merchandise consumers (Shank, 2002).

PAGE 60

60 Sample Size. To determine the sample size to test the proposed model, four factors were considered as suggested by Hair, Anderson, Ta tham, and Black (1998; pp. 604-605): (1) model misspecification, (2) model size, (3) departures from normality, and (4) estimation procedure. Hair et al. suggested increasing the sample size if specification error is suspected or if that potential construct(s) might have been omitted from the model. However, the problem may be negligible since the proposed model is based on mu ltiple existing theories in the literature (Hair et al.). Hair et al. sugge sted that the absolute minimum sample size needed to be greater than the number of covariances or correlations in the da ta matrix. Hair et al. suggested the ratio of 15 respondents for each parameter is necessary if the data is susp ected to violate multivariate normality assumption. Considering the number of potential items, a sample size of approximately 100 was needed for preliminary study, and approximately 750 was desired for the main study due to splitting the overall sample into three groups (i.e., hi gh personal involvement group, medium personal involvement group, and low personal involvement group), for the invariance analysis. Twenty one questionnaires were incomplete and thus eliminated from data analysis. The analyses were conducte d on the final sample of 736. Sampling and Data Collection Procedures A convenience sam pling method was used to surv ey the subjects from sport and fitness classes (e.g., tennis, bowling, and conditioning) in a large southern university. Institutional Review Board (IRB) approval to conduct the research was obtained from the designated institution (Protocol # 2008-U-0276). Prior to dist ributing the quest ionnaires, brief instructions were given to the respondents about the natu re of the study, includi ng purpose of the study, voluntary participation, and confiden tiality of the information to be provided. It was noted that participation in the study was voluntary and woul d in no way affect students grades in the respective courses. The respondent s were assured that the inform ation they provided would be

PAGE 61

61 anonymous. The questionnaires were distributed in a group format, although each student worked independently. During the data collection pro cedure, the responden ts were asked to answer each item included in the booklet. The time to complete the booklet was approximately 15 minutes. Instrumentation Validity and Reliability Psychom etric properties of the scales were ini tially pilot tested prio r to the main study; including internal consistency, c onstruct reliability, content and fa ce validity, construct validity, and criterion-related validity. Content validity is defined as an assessment of the degree of correspondence between the items selected to co nstitute a summated scale and its conceptual definition (Hair et al., 1998, p. 88). Content validity is generally determined in terms of three aspects including domain representa tiveness, domain relevance, and domain clarity (Safrit & Wood, 1989). A panel of experts procedure determined the content validity. Construct validity is defined as the process of determini ng the degree to which a test measures the construct it was designed to measure (Safrit & W ood, 1989, p. 37). Convergent and discriminant validity are of ten measured as types of constr uct validity (see Safrit & Wood, pp. 38-39). It is worth to noting that researchers recommended SEM as a stronger method than traditional correlation based analyses of convergent validity because SEM takes measurement error into account (Bag ozzi & Warshaw, 1990; Campbell & Fiske, 1959). Criterion validity is defined as the degr ee of correspondence between a measure and a criterion variable (Bollen, 1989). Various behaviors have typically been used as criterion variables to establish criterion-related validit y. Correlation coefficients between independent variables and criterion measures (e.g., purchase behaviors) i ndicate criterion-re lated validity (Bollen, 1989).

PAGE 62

62 Reliability is defined as the consistency of measurements when the testing procedure is repeated on a population of indivi duals or groups (American Educ ational Research Association, 1999, p. 25). Hair et al. (1998) i ndicated that various diagnosti c measures provide slightly different but related evidence for reliability: Cr onbachs alpha (interna l consistency), item-tototal correlation (general reliability), the aver age variance extracted (construct reliability). Item Selection and Elimination Justification Once item s were identified from various disciplines including marketing, sociology, psychology, and sport management, a panel of expert s consisting of four faculty from the areas of sport management, marketing, recreation, and tourism assessed content validity of the items. More specifically, afte r considering redundancy and relevancy in definition and importance in meaning, the total number of items was determined and pilot tested. Based on the results of the pilot test, modifications were made to improve the psychometric proper ties of the instrument (e.g., examination of Cronbachs alpha if item dele ted). The specific results are discussed in the Pilot Test Results section. Measurement Scales To m easure latent constructs, a number of s cales were chosen with varying numbers of items representing each construct. The latent cons tructs in the proposed model (Figure 6) include the following: Personal values (PV) Conservatism (CO) Ambition (AM) Hedonism (HE) Patriotism (PA) Team identification (TI) Attitude toward brand (AB) & Attitude toward product (AP) Intention to purchase (IP) Intention to purchase of product (IPP) Intention to purchase of brand (IPB)

PAGE 63

63 Past expenditure (PE) Perceived product attributes (PA) Expectation disconfirmation (ED) Satisfaction (SA) The scales were selected based on their ap propriateness to measure each construct. The scales were designed to ask about team licensed merchandise in particular. The following instructions were placed at the very beginning of, or intermittently inserted into, the questionnaire in order to direct the participants to think abou t the products they purchased for themselves. [ Instruction : The purpose of this section is to identify aspects that you think about when you purchased team licensed merchandise (e.g., sweat shirts, jackets, hats, T-shirt). Please think about the most recent piece of team licensed merchandise that you PURCHASED for yourself within the past two years.]. Then, the part icipants were asked to indicate what type of team licensed merchandise was purchased, what brand the product was, and what team the product depicted. Then each participant was asked to respond to each question item specific to the product, brand, and team that they listed. The following is a description of each scale. Personal Values Scale Four personal values subscales (i.e., Conservatis m Ambition Hedonism and Patriotism ) were chosen from Lee and Trails (2007) personal values scale. Lee and Trail developed a values scale that consisted of 17 personal values. Lee and Trail demonstrated that the aforementioned four personal values were significantly correlat ed with either cogniti ons about sport, sport behaviors, or both. In their study, cognitive meas ures included general sport fanship and team identification, whereas behavioral measures included televised s ports viewing, sport merchandise purchasing, readership of print media, game a ttendance, listenership of radio, and internet consumption. The correlations between the rest of the 13 personal values and the criterion measures were not significant and thus those values were not included in this study. The

PAGE 64

64 multidimensional scale result supported discrimina nt validity among the chosen four personal values constructs in Lee and Trails study. Becau se ambition and patriotism values had only two and three items, additional items were includ ed for the present research. Cronbachs alphas ranged from .60 to .91, and AVE values ranged from .425 to .778 in Lee and Trails study. Items are presented in Appendix A. Team Identification The three-item Team Identification Index (TI I) developed by Trail and James (2001) and cited in Trail, Fink et al. (2003) was used to measure the level of identification to wards a specific sport team. The TII has also s hown high reliability (i.e., Cro nbachs alpha > .85) across many studies (e.g., Trail, Anderson et al., 2005; Trail, Fink et al., 2003; see Appendix C). A sevenpoint Likert-type scale will be used to measur e team identification ranging from (1) strongly agree to (7) strongly disagree. Item s are presented in Appendix B. Attitude toward Brand and Attitude tow ard Product Olivers (1981) three items were modified to measure both an individu als attitude toward the brand and toward the product in this study. Oliver indicated that general measurement units for attitude use items that reflect affect such as like-dislike, good-bad, a nd desirable-undesirable. Thus, in the current study, Olivers three items were modified to measure attitude toward the brand that each respondent iden tified, while another set of thr ee items were further used to measure attitude toward the product (team license d merchandise that the respondent purchased). One item was created and added (i.e., My feelings toward the BRAND/team licensed merchandise I purchased are positive). A 7-point Likert-type scale ranging from (1) strongly disagree to (7) strongly agree wa s used instead of the semantic di fferential scale. Oliver claimed reliability and validity for the scale in the text but did not report any values. This scale was

PAGE 65

65 internally consistent in other studies (e.g., = .85 in Hagger, Chatzi sarantis, & Biddle, 2001). Items are presented in the Appendix C. Intention to Purchase Hagger et al.s (2001) study provided a good exam ple of behavioral intention m easures. For example, they used items such as I intend to purchase a PROD UCT A, Purchasing a PRODUCT A is something I plan to do, and I will try to purchase a PRODUCT A. Items were slightly be modified by adding In th e future as a prefix, and PRODUCT A was replaced with team licensed merchandise (e.g., In the future, purchasing team licensed merchandise is something I plan to do). In a ddition, the same set of items were created to measure intention to purchase the same brand to which each participant referred at the beginning of the survey (e.g., In the future, I intend to pu rchase more team licensed merchandise of the same brand). These items were anchored by a 7-point Likert-type scale ranging from (1) strongly disagree to (7) strongly ag ree. This scale was internally consistent in Hagger et al.s study ( = .77). Items are presented in Appendix D. Past Product Expenditure A screening question was asked of the responde nts to determ ine whether they had ever purchased a piece of team licensed merchandise. If a respondent had previously purchased at least one piece of team licensed merchandise over the last two years, the next question asked the magnitude of the respondents consumption (i.e., dollar amount spent over the past two years). Items are presented in the Appendix E. Expectancy (Dis)Confirmation Trail, Anderson et al. (2005) m odified Olivers (1980, 1981) expectancy (dis)confirmation scale and used it to measure an individuals (dis )confirmation of expectations. More specifically, the items in the scale evaluated an indivi duals experience about the quality of the

PAGE 66

66 play/performance to the extent to which it was better, equal, or worse than expected. Alpha reliability was .91 and AVE value was .67 (Trail, Anderson et al.). In this study, the items measured the respondents (dis)confirmation of e xpectations about specif ic product attributes (overall quality and performance, value, attributes, and benefits) using a Likert-type scale ranging from (1) much worse than expected, (4) as expected, to (7) much better than expected. Items are presented in the Appendix F. Satisfaction The respondents were asked to express their general level of satisfaction with a previously purchased product. More specifically, the item s measured the respondents satisfaction with decision/choice to buy and quality/value of the pr oduct. Olivers (1980) satisfaction items with flu shot experience were modified to be mo re relevant to sport merchandise purchase. Westbrook and Oliver (1981) used a multitrait-multimethod technique to determine that both a semantic differential scale and a Likert-type scale on sa tisfaction items had higher reliabilities and validities than other type s of response formats. Cronbachs coefficient alpha was high for both scales (above .91 and .75, respectively). Consideri ng these results, a 7-po int Likert-type scale ranging from (1) strongly disagree to (7) strongl y agree was used. Items are presented in the Appendix G. Perceived Product Attributes To m easure perceived value of product attributes, Netemeyer, Krishnan, Pullig, Wang, Yagci, Dean, Ricks, and Wirths (2004) perceive d quality(PQ)/perceived value for cost (PVC) items were used. Four of the eight original items were slightly modifie d. Three items (#1, #2, and #3) represent PVC whereas two items (#4 and #5 ) represent PQ. Two items were not selected because of redundancy (both items 1 and 4 in the original study measured quality), and one item using a vague description of perceived quali ty (i.e., best brand in its product class) was

PAGE 67

67 not selected. One item measuring the aesthetic attributes of products was added. The original items were reliable and valid (Cronbachs al pha = .87 .96, correlations among the items ranged .41 to .79 indicating discriminant validity, and all AVE values we re greater than .50 indicating construct reliability). Seven-point Likert-type scale ranging from (1) strongly disagree to (7) strongly agree was used. Items ar e presented in the Appendix H. Personal Involvement Zaichkowskys (1994) involvem ent with produc t scale was used. Among the original 10 items, four items were selected that measure degree of products rele vance to the subject, meaningfulness to the subject, valuableness to the subject, and interest to the subject. The original scale had the followi ng properties: Content validity checked (p. 61), test-retest reliabilities ranged from 73-.84 (pp. 61-62), and Cronbachs alpha ranged from .91-.96 (pp. 6162). Items are presented in the Appendix I. Data Analysis To statistically exam ine the structural relati onships among the latent constructs within the model, we conducted correlation analysis, confirma tory factor analysis (CFA), and structural equation modeling (SEM). Using EQS, the robust maximum likelihood (ML) method using a direct estimation process was employed to estim ate the hypothesized model. Both kurtosis and skewness were examined based on Mardias coefficient of multivariate kurtosis to detect any violation of assumptions (e.g., nonnormaility of data). Examination of multivariate kurtosis (Mardias coefficient ranged from 104.07 to 395.25 and normalized estimate ranged from 12.56 to 73.32) indicated that chi-square might be overestimated and fit indices and standard error of parameters might be underestimated (Hoyle, 1995) Thus, the Satorra-Bentle rs chi-square (S-B robust root mean square error of approximation (*RMSEA), robust comparative fit index (*CFI), the standardized root mean square resi dual (SRMR), and the chisquare per degree of

PAGE 68

68 freedom value ( df ) were reported. According to Hu, Bentler, and Kano (1992), S-B is the most reliable test statistic when evalua ting covariance structure models under various distributions and sample sizes. A multi-group (MG) invariance test allows re searchers to compare model(s) across groups. Byrne (2006) suggested that after baseline models are established, th e following sets of parameters are most commonly of interest in answering questions related to MG invariance test: (1) testing for invariance in fact or-loading paths, (2) te sting for invariance in factor covariances, and (3) testing for invariance in structural regression paths. To ascertain distinction between HIand LOW-personal involvement groups, a t test was conducted. To establish a baseline measurement model, confirmatory factor analys es (CFAs) were initially conducted using the total sample and then separate structural eq uation modeling procedures were conducted using separate sample (HIand LOW-PI subjects). For the evidence of statistical differences across groups, typically, tests of invariance are based on the chi-square difference test or the Lagrange Multiplier Test (LMTest) in the EQS program. Both tests were conducted in the current study. More specifically, a chi-squa re difference test was conducte d between the model that no constraints were imposed among the parameters a nd the model that had equality constraints on all structural paths. In additi on, MG invariance is also confirme d if the LMTest does not identify any paths that need to be free d, indicating good fit between the m odel and the data. In assessing results from the test at the univari ate level, the primary interest is the probability value associated with the LMTest statistic assigned to each equally cons trained parameter. The probability value equal to or less than .05 is an indicati on of factorial noninvari ance (Byrne & Watkins, 2003).

PAGE 69

69 CHAPTER 4 RESULTS Using EQS, the hypothesized relationships were exam ined among Personal Values (Conservatism, Hedonism, Patriotism, and Amb ition), Team Identification, Attitude toward Brand, Attitude toward Product, Intention to Purchase (intenti on to purchase of product and brand), Previous experience, Perceived Product Attributes, Expectancy Disconfirmation, and Satisfaction (the structural model is presented in Figure 1). In the model, absences of a path connecting variables imply lack of hypothesized direct effect. Results of various data analyses (i.e., CFA, SEM, Multi-group invariance test) are presented in the following sections. Pilot Test Results During the p reliminary phase, items measuring each construct were pilot tested with a convenience sample of 80 undergraduate students in a sports marketing class. The examination of reliability scores (i.e., Cronbachs alpha, item-to-total correlation, and Cronbachs alpha if item deleted) were calculated using SPSS 13.0. Cronbachs alpha ranged from .690 to .878, itemto-total correlation ranged from .327 to .803, an d Cronbachs alpha if item deleted ranged from .533 to .868. Item-to-total correlations were clos e to or above the recommended value of .50 (Hair et al., 1998) in all subscales except four (i.e., one in hedonism, one in ambition, and two in perceived product attributes). Based on the results of this pilot test, modifi cations were made to improve the internal consistency of the subscales. Mo re specifically, to reduce the numb er of measurement items, five items that contributed to reliability the leas t were eliminated (i.e., ED2, CO2, AM4, SA2, and PA5). More specifically, the am bition item with the lowest item-to-correlation (.346) was removed from the original model but a hedonism item with a lower item-to-correlation (.344) was kept because this construct had only three items. The subscale mean values were used to

PAGE 70

70 calculate correlations among th e subscales, and the highest correlation was .772 (SA-AP), supporting discriminant validity (Kline, 2005). Ba sed on the overall findings, the scales were deemed to be internally consistent (see Appendix). Main Study Results Normality Check. The assum ptions of multivariate normality and linearity were evaluated through descriptive statistics using SPSS and CFA using EQS 6.1. The analyses were performed on the total sample. Skewness a nd kurtosis values for the 36 ma nifest variables ranged from |0.08| to |1.47| a nd |0.02| to |1.88|, respec tively, which are within th e range of two standard deviations (SPSS, program manual). The freque ncy distribution of the residual covariances appeared to be symmetric (97.60% of the residuals centered around zero). Psychometric Properties of the Scales Reliabilities (Construct, Internal Cons istency, and Item-T o-Total Correlation) In the main study, using the total sample, the CFA indicated that average variance extracted (AVE) values reached the .50 cut-off level (Fornell & Larcker, 1981; Hair et al., 1998) in 7 out of 9 latent variables, except hedoni sm (.443) and ambition (.399). Cronbachs alpha for the latent variables ranged fr om .71 to .90, and item-to-total correlations ranged from .273 to .832. Overall factor loadings, AVE values, Cronb achs alphas, and item-to-total correlation values met the suggested criteria, except ambition, while hedonism met all of the criteria except for the AVE values (see Table 1). Discriminant Validity Initial discrim inant validity was evidenced in that 34 out of 36 correlations among the latent variables in the hypot hesized model were lower th an .850 (Kline, 2005). The two exceptions were the correlations between IPB AB and SAAP. However, 4 out of 9 latent variables (i.e., IPB, AB, PA, and SA) failed Fornell and Larckers (1981) more stringent test of

PAGE 71

71 discriminant validity (i.e., discriminant validity is evident when the squared correlations between one construct and any others are lower than the AVE for each construct; see Table 2). Group Classification To explore whether factoria l invariance existed am ong the parameters in the proposed model between personal involvement (PI) groups, the total sample ( N = 736) was divided using mean scores from the personal involvement subs cale. The total sample was divided into three groups (high-, medium-, and low-personal invo lvement groups). Two hundred and twenty one (30.0%) rated equal to or below 3.75 in PI and we re classified as the low personal involvement (LOW-PI) group, whereas 243 (33.0%) rated equal to or above 5.0 in PI and were classified as the high personal involvemen t (HI-PI) group in this study. Two hundred and seventy two subjects (37.0%) who rated medium values (between 4.00 or 4.75, inclusive) in personal involvement were not included for further data analysis. Exclusion of medium personal involvement subjects was to ascertain the di stinction between HIand LOW-PI groups. The results of the t test indicated that the tw o groups were significantly di fferent from each other ( t = 22.70, p < .001; refer to Table 4). Baseline Model Establishment Establishm ent of a baseline model was performe d in nine phases: step 0 through 5. In step 0, a CFA of the original measurement model th at had 12 first-order latent variables was conducted. In step 1, a CFA of a modified model (b aseline model) that had 9 first-order latent variables was conducted. In step 2, an SEM of the baseline model was conducted using the total sample. In steps 3a and 3b, SEMs of the baseli ne model were conducted using separate samples (HIvs. LOW-PI groups). In step 4, a simultane ous CFA was conducted using a combined set of both the HIand LOW-PI groups. In step 5, a simultaneous SEM was conducted using the

PAGE 72

72 combined set of both the HIand LOW-PI groups. Detailed results are presented in the following sections. Step 0. Confirmatory factor analys is (CFA) was used to estimate the measurement model using the total sample ( N = 736). The CFA was conducted for th e 12 first-order latent factors (PVCO, PVHE, PVPA, PVAM, TI, AB, AP, IPB, I PP, PA, SA, and ED). The results indicated that the measurement model fit the data well. G oodness-of-fit statistics were as follows (see Table 5): [S-B 2 (968) 2861.25; *RMSEA = .056 (90% CI for *RMSEA = .053, .058), *CFI = .877, SRMR = .068, and 2/ df = 2.96]. Based on the initial CFA results and various reliability assessments (i.e., Cronbachs alpha, alpha if it em deleted, and item-to-total correlation), both PVCO and AP were dropped from the model. Mo re specifically, PVCO was dropped due to low reliability, while AP was dropped due to multicolli nearity. In addition, because of the conceptual link between AP and IPP, IPP was also dropped. More specifically, because the original model (see Figure 6) had two attitudes (AB and AP), two facets of IP were considered and added into the model (i.e., IPP and IPB). In addition, IP was represented by two first-order factors, IPP and IPB, and was conceptually linked to AP. Due to this conceptual link between AP and IPP, elimination of IPP when AP was dropped would make sense. Step 1. The modified model from step 0 was furt her tested using CFA. As a result, model fit was slightly improved, and this became the baseline model. Goodness-of-fit statistics were as follows (see Table 5): [S-B 2 (558) 1649.99; *RMSEA = .055 (90% CI for *RMSEA = .052, .058), *CFI = .908, SRMR = .046, and 2/ df = 2.96]. Step 2, 3a, and 3b. Because the measurement model fit the data well, we were confident in proceeding to the next step, examination of the structural model. Three phases of SEM of the baseline model were subsequently performed: a) SEM using the total sample, b) SEM using the

PAGE 73

73 HI-PI group, and c) SEM using the LOW-PI group. Results from these analyses yielded a reasonably-fitting to a well-fitti ng model to each data set. Goodne ss-of-fit statistics were as follows (see Table 5): [SEM using total sample: S-B 2 (614) 1996.66; *RMSEA = .062 (90% CI for *RMSEA = .059, .065), *CFI = .874, SRMR = .183, and 2/ df = 3.25], [SEM using HI-PI group: S-B 2 (614) 962.94; *RMSEA = .053 (90% CI for *RMSEA = .047, .060), *CFI = .891, SRMR = .118, and 2/ df = 1.57], and [SEM usi ng LOW-PI group: S-B 2 (614) 1117.21; *RMSEA = .070 (90% CI for *RMSEA = .063, .076), *CFI = .801, SRMR = .139, and 2/ df = 1.82]. Multi-group Measurement Invariance Test (simultaneous CFA) Step 4a and 4b. After the baseline model was established, a test of equality in the m easurement model across groups was perfor med. Personal involvement was used as a categorical variable (model 1 = hi-persona l involvement and model 2 = low-personal involvement) to perform the test. Then, the CFA for the baseline model was conducted simultaneously for both groups. The factor loadin gs and covariates for HIand LOW-PI groups were compared to determine whether the constr ucts were equally measured between two groups. Initially, a simultaneous CFA without the factor loadings and covariates for HIand LOW-PI groups constrained to be equal was conducted. The GFIs for the unconstrained model were: SB 2 (1136) 2245.71; *RMSEA = .070 (90% CI for *RMSEA = .065, .074), *CFI = .820, SRMR = .286, and 2/ df = 1.98. A subsequent simultaneous CFA wa s conducted with the factor loadings and covariates for HIand LOW-PI groups constrai ned to be equal. Equality constraints were imposed for 63 parameters to be tested (refer to Tables 6 and 7). The GFIs for the constrained model were: S-B 2 (1197) 2129.69; *RMSEA = .062 (90% CI for *RMSEA = .058, .066), *CFI = .848, SRMR = .350, and 2/ df = 1.78. Then, chi-square differe nce test was conducted between the two models, and the results indicated that difference was significant [The chi-square difference value was 116.02(61); p < .05. However, due to a large samp le size, it is possible that

PAGE 74

74 Delta chi-square might have been inflated, indicating a lack of difference. Some overlap in confidence interval of RMSEA indeed provided a contrast result suppor ting the non-significant difference between the two groups; .065 ~ .074 and .058 ~ .066]. MG invariance is also confirmed in that the LMTest did not identify a ny paths that need to be freed, indicating good fit between the model and the data. At the univariat e level, the probability values for 27 factor loadings for HIand LOW-PI groups constrained to be equal were .410 or greater, indicating factorial invariance (refer to Table 6). The proba bility values for 36 cova riates for HIand LOWPI groups constrained to be equa l were .650 or greater, indicating factorial invariance (refer to Table 7). Thus, the overall results indicate that the model with equal loadings for HIand LOWPI groups fits as well as the model without the equality constraints. Multi-group Structural Invaria nce Test (simultaneous SEM) Step 5a and 5b. To determ ine if (non)invariance in regression paths between HIand LOW-PI group existed, two phases of SEMs of the baseline model were performed simultaneously for both groups: model 1 = HI-PI and model 2 = LOW-PI. Thus, the structural (regression) paths for HIand LOW-PI groups were compared to determine whether the structural relationships did change due to va rying degree of personal involvement. Initially, a simultaneous SEM without the regression coefficien ts for HIand LOW-PI groups constrained to be equal. All 15 structural paths depicted in Figure 9 were constrained to be equal between HIand LOW-PI groups. The GFIs for the unconstrained model were: S-B 2 (1228) 2025.37; *RMSEA = .059 (90% CI for *RMSEA = .054, .063), *CFI = .868, SRMR = .118, and 2/ df = 1.65. Another simultaneous SEM was conducted with the regression coefficients for HIand LOW-PI groups constrained to be equal. The GF Is for the constrained model were: S-B 2 (1243) 2025.28; *RMSEA = .058 (90% CI for RMSEA = .053, .062), *CFI = .871, SRMR = .118, and 2/ df = 1.63. Then, chi-square difference test was conduct ed between the two m odels, and the results

PAGE 75

75 indicated that difference was small [T he chi-square difference value was 0.09(15); p < .05; a large overlap in confidence interval of RMSEA also supports the non-significa nt difference between the two groups; .054 ~ .063 and .053 ~ .062]. Thus, the overall results indicate that the model with equal path coefficients for HIand LOW-PI groups fits as well as the model without the equality constraint. While the goodness-of-fit indices indicted that structural invariance may exist across groups, each of the equally constrained regression pa th was further tested at the univariate level. The probability values of the univariate incremen t for all constraints indicated that all the parameters were invariant between the two groups. More specifica lly, the probabili ty values for 15 structural (regression) paths for HIand LOWPI groups constrained to be equal were .579 or greater, indicating factorial i nvariance (refer to Table 8).

PAGE 76

76 Table 4-1. Means (M), standard devi ations (SD), beta coefficients ( ), internal consistency ( ), and average variance extracted (AVE) for the scales ( N = 736) Item M SD ra AVE Personal values Hedonism Self-indulgence is an important value to me Sensuous gratification is an important value to me Hedonism is an important value to me Intense pleasure is an important value to me Patriotism Nationalism is an important value to me Loyalty to country is an important value to me Patriotism is an important value to me Devotion to my country is an important value to me Ambition Ambition is an important value to me Competition is an important value to me Aspiring to greatness is an important value to me Aggressively pursuing my goals is an important value to me 4.07 4.69 3.83 4.61 5.16 5.49 5.31 5.36 6.05 5.52 5.98 5.89 1.45 1.15 1.34 1.49 1.32 1.30 1.31 1.35 1.00 1.21 1.04 1.12 .670 .706 .555 .718 .717 .892 .886 .899 .563 .576 .761 .607 .449 .498 .308 .515 .513 .795 .784 .808 .317 .332 .579 .369 .75 .90 .71 .443 .725 .399 Team identification I consider myself to be a real fan of the team I would experience a loss if I had to stop being a fan of the team Being a fan of the team is very important to me 5.98 5.12 5.45 1.36 1.72 1.42 .768 .824 .910 .590 .679 .828 .86.699 Attitude toward brand I like the brand I think that the brand is good I think that the brand is desirable My feelings toward the brand are positive 4.83 5.00 4.97 5.10 1.40 1.22 1.22 1.16 .769 .839 .871 .865 .591 .703 .491 .748 .90.700 Intention to purchase of brand In the future, I intend to purchase more of the same brand of team licensed merchandise In the future, I am likely to purchase more team licensed merchandise of the same brand In the future, I intend to purchase more team licensed merchandise of the same brand 4.86 5.02 5.23 1.31 1.29 1.32 .866 .888 .813 .749 .789 .661 .89 .733 Perceived product attributes In general, what I get from team licensed merchandise is worth the cost All things considered (price, time, and effort), team licensed merchandise is a good buy In general, compared to other products, team licensed merchandise is a good value for the money In general, team licensed mercha ndise consistently performs better than other products In general, compared to other pr oducts, team licensed merchandise is of very high quality In general, team licensed merchandise is aesthetically pleasing 4.79 4.95 4.42 4.25 4.53 5.11 1.32 1.17 1.32 1.24 1.21 1.23 .709 .765 .744 .680 .739 .720 .503 .586 .553 .463 .546 .519 .87 .528 Note. Statistical method: confirmatory factor anal ysis (EQS). a. Item -to-total correlation.

PAGE 77

77 Table 4-1. Continued Item M SD ra AVE Expectancy disconfirmation The overall quality of the team licensed merchandise was The overall performance of the team licensed merchandise was The attributes of the team licensed merchandise were The benefits of the team licensed merchandise were 4.80 4.82 4.79 4.72 0.93 0.95 0.93 0.91 .677 .660 .863 .852 .458 .435 .745 .725 .86.634 Satisfaction I was satisfied with my decision to buy the team licensed merchandise I was satisfied with the value of the team licensed merchandise My choice to buy the team licensed merchandise was a wise one I think that I did the right thing when I decided to buy the team licensed merchandise 5.84 5.29 5.27 5.30 0.96 1.14 1.15 1.12 .712 .771 .856 .837 .507 .594 .733 .700 .87.591 Note. Statistical method: confirmatory factor anal ysis (EQS). a. Item -to-total correlation.

PAGE 78

78Table 4-2. Correlations among latent vari ables of the original model: step 0 ( N = 736) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Conservatism (1) 1 Hedonism (2) .643* 1 Patriotism (3) .415 .234 1 Ambition (4) .279 .427 .537 1 Team identification (5) .375 .292 .560 .487 1 Attitude toward brand (6) .406 .374 .511 .506 .376 1 Attitude toward product (7) .443 .427 .539 .575 .632 .732 1 Intention to purchase of product (8) .411 .330 .505 .505 .784 .541 .867** 1 Intention to purchase of brand (9) .484 .422 .530 .525 .544 .880** .819* .738 1 Perceived product Attributes (10) .551 .459 .479 .512 .555 .742* .896** .721 .828* 1 Satisfaction (11) .486 .459 .511 .591 .637 .703 1.02** .857** .805* .922** 1 Expectation Disconfirmation (12) .373 .408 .293 .407 .334 .548 .685 .487 .583 .717 .685 1 Note Statistical method: confirmatory factor analysis (EQS). Correlation failed Forne ll and Larckers (1981) stringent test of discriminant validity. ** Correlati on failed for both Klines (2005) initial test of discriminant validity and Fornell and Larc kers (1981) stringent test of discriminant validity.

PAGE 79

79Table 4-3. Correlations among latent vari ables of the baseline model: step 1 ( N = 736) (1) (2) (3) (4) (5) (6) (7) (8) (9) Hedonism (1) 1 Patriotism (2) .249 1 Ambition (3) .427 .540 1 Team identification (4) .307 .561 .481 1 Attitude toward brand (5) .378 .510 .514 .389 1 Intention to purchase of brand (6) .431 .535 .532 .549 .881** 1 Product attributes (7) .469 .483 .519 .564 .746 .831 1 Satisfaction (8) .473 .513 .582 .639 .699 .806* .919** 1 Expectancy disc onfirmation (9) .410 .296 .411 .343 .555 .588 .713 .684 1 Note Statistical method: confirmatory factor analysis (EQS). Correlation failed Forne ll and Larckers (1981) stringent test of discriminant validity. ** Correlati on failed for both Klines (2005) initial test of discriminant validity and Fornell and Larc kers (1981) stringent test of discriminant validity.

PAGE 80

80 Table 4-4. T-test between HIand LOW-personal involvement groups t df Sig. (2-tailed) Variable Personal involvement 22.70 458 p < .001 Note LOW-PI group: N = 216, M = 2.99, SD = .71, and SE = .05; and HI-PI group: N = 243, M = 5.68, SD = .57, and SE = .04.

PAGE 81

81Table 4-5. Summary of fit sta tistics for testing multi-group in variance of the proposed model Procedures S-B df) RMSEA(CI) CFI SRMR Outcome interpretation Step 0 Measurement test of orig inal model: CFA using total sample no constraints imposed 2861.25(968) .056(.053; .058)* .877* .068 Measurement model fit the total sample well (Factors included in the model: PVCO, PVHE, PVPA, PVAM, TI, AB, AP, IPB, IPP, PA, SA, & ED) Step 1 Measurement test of modified (baseline) model: CFA using total sample no constraints imposed 1649.99(558) .055(.052; .058)* .908* .046 Measurement model of the modified model fit the total sample well (Factors included in the model: PVCO, AP, IPP, IP were dropped from the original model) Step 2 Structural test of baseline model : SEM using total sample no constraints imposed 1996.66(614) .062(.059; .065)* .874* .183 Structural model fits the data well Step 3a Structural test of baseline model: SEM using hi personal involvement group no constraints imposed 962.94(614) .053(.047; .060)* .891* .118 Structural model fits the hi personal involvement group well Step 3b Structural test of baseline model: SEM using low personal involvement group no constraints imposed 1117.21(614) .070(.063; .076)* .801* .139 Structural model reasonably fit the low personal involvement group Note Fit indices are suggested to meet the following criteria for i nvariance across groups: CFI close to or greater than .90 (Byrn e & Campbell, 1999), RMSEA < .06 (Hu & Bentler, 1999) SRMR < .08 (Hu & Bentler, 1999), and S-B 2/ df < 3.0 (Kline, 2005). (Results are based on Satorra-Bentlers robust statistics).

PAGE 82

82Table 4-5. Continued Procedures S-B df) RMSEA(CI) CFI SRMR Outcome interpretation Step 4a Measurement invariance test: simultaneous CFA no constraints imposed 2245.71(1136) .070(.065; .074)* .820* .286 Step 4b Measurement invariance test: simultaneous CFA invariant factor loadings and factor covariances constrained equal between groups 2129.69(1197) .062(.058; .066)* .848* .350 Measurement equality exists between hiand low-personal involvement groups Step 5a Structural invariance test: simultaneous SEM no constraints imposed 2025.37(1228) .059(.054; .063)* .868* .118 Step 5b Structural invariance test: simultaneous SEM invariant regression paths and factor variancecovariances constrained equal between groups 2025.28(1243) .058(.053; .062)* .871* .118 Structural inequality exists between hiand low-personal involvement groups Note Fit indices are suggested to meet the following criteria for i nvariance across groups: CFI close to or greater than .90 (Byrn e & Campbell, 1999), RMSEA < .06 (Hu & Bentler, 1999) SRMR < .08 (Hu & Bentler, 1999), and S-B 2/ df < 3.0 (Kline, 2005). (Results are based on Satorra-Bentlers robust statistics).

PAGE 83

83Table 4-6. Multivariate lagrange multiplier test by simulta neous CFA (invariance in factor loadings from step 4) Cumulative multivariate statistics Univariate increment Step Constraint Parametera df Probability Probability* 1 Constraint: 27 (1, ED4, ED) = (2, ED4, ED) .679 1 .410 .679 .410 2 Constraint: 11 (1, TI3, TI) = (2, TI3, TI) .997 2 .607 .319 .572 3 Constraint: 7 (1, PVAM2, AM) = (2, PVAM2, AM) 1.345 3 .718 .348 .555 5 Constraint: 8 (1, PVAM3, AM) = (2, PVAM3, AM) 1.739 5 .884 .187 .665 7 Constraint: 20 (1, PA5, PA) = (2, PA5, PA) 2.123 7 .953 .183 .669 8 Constraint: 13 (1, AB3, AB) = (2, AB3, AB) 2.293 8 .971 .170 .680 16 Constraint: 18 (1, PA3, PA) = (2, PA3, PA) 3.131 16 1.000 .080 .778 17 Constraint: 21 (1, PA6, PA) = (2, PA6, PA) 3.213 17 1.000 .082 .774 18 Constraint: 1 (1, PVHE2, HE) = (2, PVHE2, HE) 3.300 18 1.000 .087 .768 19 Constraint: 19 (1, PA4, PA) = (2, PA4, PA) 3.405 19 1.000 .105 .746 20 Constraint: 23 (1, SA3, SA) = (2, SA3, SA) 3.489 20 1.000 .085 .771 21 Constraint: 2 (1, PVHE3, HE) = (2, PVHE3, HE) 3.610 21 1.000 .121 .728 25 Constraint: 25 (1, ED2, ED) = (2, ED2, ED) 3.876 25 1.000 .054 .816 26 Constraint: 10 (1, TI2, TI) = (2, TI2, TI) 3.944 26 1.000 .069 .793 27 Constraint: 26 (1, ED3, ED) = (2, ED3, ED) 3.996 27 1.000 .052 .819 32 Constraint: 16 (1, IPB3, IPB) = (2, IPB3, IPB) 4.218 32 1.000 .044 .834 41 Constraint: 15 (1, IPB2, IPB) = (2, IPB2, IPB) 4.628 41 1.000 .043 .836 42 Constraint: 24 (1, SA4, SA) = (2, SA4, SA) 4.662 42 1.000 .033 .855 43 Constraint: 17 (1, PA2, PA) = (2, PA2, PA) 4.699 43 1.000 .038 .846 46 Constraint: 3 (1, PVHE4, HE) = (2, PVHE4, HE) 4.766 46 1.000 .018 .893 49 Constraint: 12 (1, AB2, AB) = (2, AB2, AB) 4.794 49 1.000 .007 .932 50 Constraint: 6 (1, PVPA4, PVPA) = (2 PVPA4, PVPA) 4.799 50 1.000 .005 .943 51 Constraint: 14 (1, AB4, AB) = (2, AB4, AB) 4.803 51 1.000 .004 .950 57 Constraint: 9 (1, PVAM4, AM) = (2, PVAM4, AM) 4.823 57 1.000 .001 .973 58 Constraint: 4 (1, PVPA2, PVPA) = (2 PVPA2, PVPA) 4.823 58 1.000 .001 .980 59 Constraint: 22 (1, SA2, SA) = (2, SA2, SA) 4.824 59 1.000 .001 .981 62 Constraint: 5 (1, PVPA3, PVPA) = (2 PVPA3, PVPA) 4.826 62 1.000 .001 .977 Note (Probability values for univa riate increment equal to or less than .05 indica te structural noninvarian ce in each regression path; Byrne, 2006). a indicates the followings: (group1, manifest variable, factor) = (gr oup2, manifest variable, factor).

PAGE 84

84Table 4-7. Multivariate lagrange multiplier test by simultane ous CFA (invariance in covariates from step 4) Cumulative multivariate statistics Univariate increment Step Constraint Parametera df Probability Probability* 4 Constraint: 49 (1, TI, AB) = (2, TI, AB) 1.552 4 .817 .206 .650 6 Constraint: 30 (1, HE, TI) = (2, HE, TI) 1.940 6 .925 .201 .654 9 Constraint: 52 (1, TI, SA) = (2, TI, SA) 2.402 9 .983 .109 .741 10 Constraint: 59 (1, IPB, SA) = (2, IPB, SA) 2.514 10 .991 .111 .738 11 Constraint: 62 (1, PA, ED) = (2, PA, ED) 2.647 11 .995 .133 .715 12 Constraint: 33 (1, HE, PA) = (2, HE, PA) 2.757 12 .997 .110 .740 13 Constraint: 41 (1, PVPA, SA) = (2, PVPA, SA) 2.847 13 .998 .090 .764 14 Constraint: 43 (1, AM, TI) = (2, AM, TI) 2.961 14 .999 .114 .736 15 Constraint: 58 (1, IPB, PA) = (2, IPB, PA) 3.051 15 1.000 .090 .764 22 Constraint: 35 (1, HE, ED) = (2, HE, ED) 3.679 22 1.000 .069 .793 23 Constraint: 55 (1, AB, PA) = (2, AB, PA) 3.753 23 1.000 .074 .785 24 Constraint: 36 (1, PVPA, AM) = (2, PVPA AM) 3.821 24 1.000 .068 .794 28 Constraint: 54 (1, AB, IPB) = (2, AB, IPB) 4.041 28 1.000 .045 .833 29 Constraint: 38 (1, PVPA, AB) = (2, PVPA, AB) 4.095 29 1.000 .054 .817 30 Constraint: 34 (1, HE, SA) = (2, HE, SA) 4.135 30 1.000 .040 .842 31 Constraint: 50 (1, TI, IPB) = (2, TI, IPB) 4.174 31 1.000 .040 .842 33 Constraint: 56 (1, AB, SA) = (2, AB, SA) 4.260 33 1.000 .042 .838 34 Constraint: 46 (1, AM, PA) = (2, AM PA) 4.317 34 1.000 .057 .812 35 Constraint: 48 (1, AM, ED) = (2, AM, ED) 4.385 35 1.000 .068 .794 36 Constraint: 39 (1, PVPA, IPB) = (2, PVPA, IPB) 4.437 36 1.000 .052 .820 37 Constraint: 61 (1, PA, SA) = (2, PA, SA) 4.481 37 1.000 .044 .834 38 Constraint: 31 (1, HE, AB) = (2, HE, AB) 4.515 38 1.000 .034 .853 39 Constraint: 45 (1, AM, IPB) = (2, AM, IPB) 4.550 39 1.000 .036 .850 40 Constraint: 44 (1, AM, AB) = (2, AM, AB) 4.585 40 1.000 .035 .852 44 Constraint: 28 (1, HE, PVPA) = (2, HE, PVPA) 4.726 44 1.000 .027 .870 45 Constraint: 42 (1, PVPA, ED) = (2, PVPA, ED) 4.748 45 1.000 .022 .883 47 Constraint: 53 (1, TI, ED) = (2, TI, ED) 4.779 47 1.000 .013 .908 48 Constraint: 63 (1, SA, ED) = (2, SA, ED) 4.786 48 1.000 .007 .932 Note (Probability values for univa riate increment equal to or less than .05 indica te structural noninvarian ce in each regression path; Byrne, 2006). a indicates the followings: (group1, factor factor) = (group2, factor, factor).

PAGE 85

85Table 4-7. Continued Cumulative multivariate statistics Univariate increment Step Constraint Parametera df Probability Probability* 52 Constraint: 29 (1, HE, AM) = (2, HE, AM) 4.806 52 1.000 .003 .954 53 Constraint: 32 (1, HE, IPB) = (2, HE, IPB) 4.817 53 1.000 .010 .919 54 Constraint: 47 (1, AM, SA) = (2, AM, SA) 4.819 54 1.000 .002 .963 55 Constraint: 57 (1, AB, ED) = (2, AB, ED) 4.820 55 1.000 .002 .968 56 Constraint: 51 (1, TI, PA) = (2, TI, PA) 4.822 56 1.000 .001 .971 60 Constraint: 37 (1, PVPA, TI) = (2, PVPA TI) 4.825 60 1.000 .001 .979 61 Constraint: 40 (1, PVPA, PA) = (2, PVPA, PA) 4.825 61 1.000 .001 .979 63 Constraint: 60 (1, IPB, ED) = (2, IPB, ED) 4.826 63 1.000 .000 .995 Note (Probability values for univariate incr ement equal to or less than .05 indicate structural noninvariance in each regression path; Byrne, 2006). a indicates the followings: (group1, factor factor) = (group2, factor, factor).

PAGE 86

86Table 4-8. Multivariate lagrange multiplier test by simultaneous SEM (invariance in regression coefficients from step 5) Cumulative multivariate statistics Univariate increment Step Constraint Parametera df Probability Probability* 1 Constraint 15 (1, SA, ED) = (2, SA, ED) .308 1 .579 .308 .579 2 Constraint 5 (1, AB, TI) = (2, AB, TI) .504 2 .777 .197 .658 3 Constraint 1 (1, HE, PV) = (2, HE, PV) .644 3 .886 .140 .708 4 Constraint 13 (1, PA, SA) = (2, PA, SA) .742 4 .946 .098 .755 5 Constraint 4 (1, TI, PV) = (2, TI, PV) .809 5 .976 .067 .796 6 Constraint 7 (1, AB, PV) = (2, AB, PV) .903 6 .989 .094 .759 7 Constraint 12 (1, PA, TI) = (2, PA, TI) .963 7 .995 .060 .807 8 Constraint 8 (1, IPB, TI) = (2, IPB, TI) .987 8 .998 .024 .876 9 Constraint 9 (1, IPB, AB) = (2, IPB, AB) 1.033 9 .999 .046 .829 10 Constraint 6 (1, AB, SA) = (2, AB, SA) 1.059 10 1.000 .026 .872 11 Constraint 2 (1, PVPA, PV) = (2 PVPA, PV) 1.067 11 1.000 .008 .928 12 Constraint 10 (1, IPB, PA) = (2, IPB, PA) 1.074 12 1.000 .007 .935 13 Constraint 11 (1, IPB, SA) = (2, IPB, SA) 1.201 13 1.000 .127 .722 14 Constraint 14 (1, PA, PE) = (2, PA, PE) 1.202 14 1.000 .001 .977 Note (Probability values for univariate incr ement equal to or less than .05 indicate structural noninvariance in each regression path; Byrne, 2006). a indicates the followings: (group1, manifest variable or factor, fact or) = (group2, manifest variab le or factor, factor).

PAGE 87

87 Figure 4-7. Path coefficients from the mode l test in step 2. [refer to Table 4-5] AB IP B PV PA SA PE ED .710 .043 .518 .025 .602 .185 .723 .188 .942 HE PV PA PV AM PV .755 .698 .679 .422 -.335 .154 TI

PAGE 88

88 Figure 4-8. Path coefficients from the model test in step 3a and 3b. [Path coefficients for high and low personal involvement groups (the fi rst path coefficients are measured by high personal involvement s ubjects and the second path co efficients are measured by low personal involvement subjects). refer to Table 4-5] AB IP B PV PA SA PE ED .486/.513 -.058/-.009 .391/.325 -.026/.096 .595/.617 .184/.213 .475/.672 .174/.125 .903/.891 HE PV PA PV AM PV .740/1.00 .653/.327 .534/.648 .277/.156 -.312/-.160 .107/.209 TI

PAGE 89

89 Figure 4-9. Path coefficients from the model test in step 5. [Path coeffi cients for high and low personal involvement groups (the first pa th coefficients are measured by high personal involvement subjects and the second path coefficients are measured by low personal involvement subjects). refer to Table 4-5] AB IP B PV PA SA PE ED .507/.517 -.047/-.049 .392/.404 -.029/-.032 .598/.597 .210/.204 .497/.478 .144/.140 .906/.908 HE PV PA PV AM PV .734/.738 .644/.662 .602/.586 .300/.314 -.335/-.338 .113/.114 TI

PAGE 90

90Table 4-9. Differences in the strength of the hypothesi zed relationships under HIand LOW-personal involvement Hypothesized directions between latent factors Stronger relationship under (high vs. low) personal involvement Status H1: Personal values Team identification Invariant Confirmed H2: Personal values Brand attitude Invariant Confirmed H3: Personal values Product attitude Invariant Eliminated H4: Team identification Brand attitude Low PI Failed H5: Team identification Product attributes Low PI Failed H6: Team identification Product attitude High PI Eliminated H7: Brand attitude Product attitude Low PI Eliminated H8: Product attitude Intention to purchase High PI Eliminated H9: Product attributes Product attitude High PI Eliminated H10: Past expenditure Product attributes High PI Failed H11: Expectancy disconfirmation Satisfaction Low PI Failed H12: Satisfaction Product attitude Low PI Eliminated H13: Satisfaction Product attributes Low PI Failed H14: Satisfaction Intention to purchase Low PI Eliminated H15: Satisfaction Brand attitude Low PI Failed New: Team identification Intention to purchase of product Invariant New: Brand attitude Intention to purchase of product Invariant New: Product attributes Intention to purchase of product Invariant New: Satisfaction Intention to purchase of product Invariant Note Statistical method: structur al equation modeling (EQS).

PAGE 91

91 CHAPTER 5 DISCUSSION Focus of the Study The prim ary objective of this study was to deve lop a model that explains consumption of licensed team merchandise (LTM). The model consisted of 12 latent factors that were derived from values theory (Rokeach, 1973; among others), identity theory (Stryker, 1968), attitude theory (Fishbein & Ajzen, 1975), satisfaction theo ry (Oliver, 1980), and se veral other specific concepts (i.e., personal involvement and perceived product attributes ) that have been publicized to influence consumer behavior. In the model, it was originally hypothesized that the latent structural relationships flow from personal valu es to attitudes (toward the product and the brand) to purchase intention (represented by both inte ntion to purchase brand and product). Further, it was hypothesized that the impact of satisfaction (i.e., expectancy disconfirmation about the purchase and satisfactio n with the purchase) and perceived product attributes (i.e., past expenditure and perceived value of product attributes) on the forma tion of attitude toward brand and product, and purchase inten tion (brand and product). The m odel was then statistically examined in terms of its structural relationships among the constructs within the model. A secondary interest of the current study was to assess (in)equality in the measurement and structural relationships among la tent constructs due to differe nces in the level of personal involvement. To fulfill the secondary objective, th e total sample was divided into three groups (HI-, MIDDLE-, and LOW-PI), but only HIand LOW-PI groups were used. Elimination of the MIDDLE-PI group was designed to ascertain the distinction be tween HIand LOW-PI groups. Once factorial invariance was determined at the measurement level, invariance of the structural model was examined. Finally, it was hoped that accomplishment of th ese objectives would

PAGE 92

92 provide relevant management and/or marketing implications and offer pr actical suggestions. Key findings are discussed in more de tail in the following sections. Overview of the Significant Findings Establishment of the Baseline Model Two phases of confir matory factor analysis (i.e., an initial CF A for the originally proposed model and another CFA for the modified model) were performed to evaluate the measurement model. The initial CFA for the original model that included 12 first-order latent factors (PVCO, PVHE, PVPA, PVAM, TI, AB, AP, IPB, IPP, PA, SA, and ED) represented the data well. However, inspection of the interfactor correla tions indicated that a few factors were highly correlated each other, with some beyond unity (i.e ., 1.0), indicating lack of discriminant validity. In addition, based on the examination of the CF A and various tests for psychometric properties of the original model, three f actors (i.e., PVCO, AP, and IPP) were eliminated from further consideration. Therefore, the originally proposed measurement model needed to be modified and a new measurement model was specified. A subs equent CFA was performed for the modified model and as a result, the model fit slightly improved (exclusion of the three factors improved *RMSEA and *CFI values by .001 and .031, respectively). CFA for the modified model that included ni ne first-order latent factors (PVHE, PVPA, PVAM, TI, AB, IPB, PA, SA, a nd ED) indicated that the meas urement model represented the data well. The three different reliability assessments (i.e., Cronbachs alpha, item-to-total correlations, and average variance extracted) jointly demonstrated that seven of the nine firstorder latent constructs in th e modified model were reliable. The two exceptions, PVHE and PVAM, did not reach suggested levels of the rigorous construct re liability value. While most of the correlations among the latent constructs me t Klines (2005) criter ion of .85, which supports discriminant validity, two of the correlations am ong the latent constructs failed both tests of

PAGE 93

93 discriminant validity (i.e., SA-AP and AB-IPB). These items need to be improved in terms of discriminant validity for future st udies. Both the original and the modified models are presented in Figure 6 and 7, respectively. Three separate phases of the SEM of the modi fied (baseline) model were subsequently performed: a) SEM using the total sample, b) SEM using the HI-PI group, and c) SEM using the LOW-PI group. Results from these analyses indicated that the m odels reasonably fit each data set. More specifically, the results from the SEM that used the total sample indicated that the model fit the data reasonably well. As a necessary step for the next phase (i.e., multi-group measurement invariance test), the baseline model was further tested for structural relationships using each HIand LOW-PI group separately. The model fit the data from the HI-PI group well and fit the data from the LOW-PI group adequately. The SEM on the total sample further indicated th at many of the structural relationships in the proposed model were supported in that the pr oposed model fit the data reasonably well. In the model, the three first-order latent variable s, PVHE, PVPA, and PVAM, were all significantly associated with the second-order latent variable PV (57% of the common ality of the first-order latent variables). PV was signifi cantly associated with TI, expl aining 50% of the variance. PV, TI, and SA were significantly associated with AB explaining 60% of the variance. TI, PE, and SA were significantly associated with PA, expl aining 89% of the variance. ED was significantly associated with SA, explaining 52% of the va riance. TI, AB, PA, and SA were significantly associated with IPB, explaining 83% of the varian ce. Comparison of both of the path coefficients from the model tested by th e HI-PI and LOW-PI groups i ndicated some variation. The discrepancies in path coefficients between the two groups ranged from .012 (SA-PA) to .326 (PVAM-PV). Although some variation appear to exist among path coefficients between the two

PAGE 94

94 groups due to differences in the level of PI, the univariate statistics of multi-group invariance test were all invariant, and thus, did not support ma ny of the secondary hypotheses (refer to Figure 9). These multiple procedures established a baseline model that measured LTM consumption behavior within the context of sport and was used to explore the factoria l invariance test across PI groups. Comparison with the Existing Literature This section discusses th eories that support the proposed rela tionships between each latent construct that are conceptually linked to each ot her. The results of the SEM (Step 2) in the current study were generally cons istent with prior studies. For instance, 12 hypotheses (original and new) were tested based on a premise that la tent constructs are conceptually linked to each other, composing the originally proposed model. Literature supported these premises in that many conceptualizations elucidated various consumption behaviors. Those include values theory (Rokeach, 1973a, 1973b), identity theory (Stryker, 1968), attitude theory (Fishbein & Ajzen, 1975), satisfaction theory (Oliver, 1980), etc. In the current study, PV explained 50% of the variance in TI. This finding supports Lee and Trails (2007) results in that they f ound ambition, patriotism, and hedonism were all significantly correlated w ith TI (correlations ranged from .12 to .40). While Lee and Trails study may be one of a few studies that have investigated the relati onship between personal values and team identification, ther e are very rare empirical findings that allow us to compare the result of the current study with the existing literature. This si gnifies the need for further study that empirically explores the relationship between PV and TI. In the current study, PV, TI, and SA explaine d 60% of the variance in AB. The influence of TI on the formation of AB partially supports Gladden and colleagues study of brand equity. More specifically, Gladden and others (Gla dden, Irwin, & Sutton, 2001; Gladden & Milne, 1999;

PAGE 95

95 Gladden, Milne, & Sutton, 1998) argued that co nsumers attitude toward a brand constitutes brand association, which is derived from the conc ept of team identification. However, the sign was negative in the current st udy indicating team identificat ion impacts brand attitude negatively. This inconsistent find ing also signifies the need for additional studies. In the current study, out of the three factors, SA contributed the mo st to the formation of AB. This finding partially supports Oliver and Lindas (1981) argument in that satisfaction (with a product, specifically sleeping apparel) mediates the re lationship between expect ancy (dis)confirmation and attitude (toward a product, ag ain, sleeping apparel) or inten tion. In Oliver and Lindas study, satisfaction explained up to 4% of the variance in attitude. The finding that SA contributes to the formation of AB is also somewhat consistent with Madrigals (2003) finding that performance satisfaction explained 15% of the variance in optimism about the teams future performance, which is an equivalent concept to attitude because it used affect to represent the optimism. TI, PE, and SA explained 89% of the variance in PA in the current study. Out of the three factors, SA contributed the most to PA, while the influence of TI was minimal at best. This finding is somewhat inconsistent with Kwon, Tr ail, and Jamess (2007) study in which they tested three models (a direct effect, a partially mediated, and a fully mediated) explaining relationships among team identif ication, perceived value of pr oduct attributes, and purchase intention. Kwon et al. indicated that team id entification influenced purchase intention of licensed-sport apparel, which was mediated by co nsumers perceived valu e of product attributes. Team identification explained a fair amount of va riance (13.2%) in percei ved value, but in the current study, the influence of TI on PA was weak. In the current study, it was also found that ED explained 52% of the variance in SA. This result is consistent with pr evious studies in that numer ous researchers have found that

PAGE 96

96 expectancy disconfirmation ofte n explained a large amount of variance in satisfaction for running competition (Caro & Garcia, 2006), genera l sport fan satisfaction (Madrigal, 1995), spectators game satisfaction (Tra il, Fink et al., 2003; Leeuwen et al., 2002), etc. Leeuwen et al. argued that disconfirmation of pr eexisting expectations is direc tly associated with customer satisfaction. The influence of ED on SA that the current study found was also consistent with Oliver and Lindas study in that ED explained a large amount of vari ance (21% and 30% for male and female consumers, respectively) in satisfaction with genera l products (i.e., sleeping apparel). In the current study, TI, AB, PA, and SA expl ained 83% of the variance in IPB. These results are somewhat consistent with previous findings in that Matsuoka et al. (2003) showed that satisfaction influenced intention to atte nd future games (satisfaction with performance explained 26% of intentions to attend future games). The influence of product attributes on purchase intention was also found in previous studies in that perceived value of product attributes explained 42.6% of the variance in purchase inten tion (Kwon, Trail et al.). The influence of satisfaction on purchas e intention was also found in prev ious studies in that attitudes explained 8% of the variance in intenti on to attend a hockey game (Cunningham & Kwon, 2003). Out of the four factors exam ined in the current study, AB contributed the most to IPB. This relationship of attitude w ith behavioral intention was also found in Madrigals (2001) study in that a hierarchical relations hip existed among beliefs, attitudes, and inten tions in relation to televised sports consumption. Overall multi-group (MG) invariance test results are discussed in the subsequent sections. Multi-Group (MG) Invariance Test The MG invarian ce test consists of two aspects: measuremen t and structural invariance (Byrne & Watkins, 2003). The measurement invari ance test involves testi ng the equivalence of

PAGE 97

97 measured constructs in two or more independent groups to assure that the same constructs are being assessed in each group (Chen, Sousa, & West, 2005, p. 472). Inspection of GFIs indicated a reasonable fit between the model and the data, suggesting invariance of the measurement constructs between HIand LOW-PI groups. Examin ation of both factor lo adings and covariates in simultaneous CFAs indicates that tests of the equivalence of the measured constructs between HIand LOW-PI groups assure that the same c onstructs are being asse ssed in each group. These results provided further evidence for the ne xt step, testing stru ctural invariance. Simultaneous SEM was then conducted to answer the following research question in the current study: due to the mode rating functions of the involvement strength of the relationships among the variables would differ for HIand LOWPI groups. Inspection of GFIs indicated a reasonable fit between the model and the data, suggesting invariance ac ross PI groups. This finding did not support the research question in general. To a ssess if any parameters were noninvariant across PI groups, an LM test was perfo rmed, and the results in dicated that all (15) structural (regression) paths were invariant between HIand LO W-PI groups. As a result, two original hypotheses were supported. More specifically, it was s upported that differences in the levels of PI did not change the strength of th e relationship between PV and TI, while the second hypothesis that differences in the levels of PI did not change th e strength of the relationship between PV and AB was also supported. However, the results of the simultaneous SEM did not support six original hypotheses that each hypothesi zed relationship would be either stronger or weaker under HIor LOW-PI. Seven original hyp otheses were eliminated due to the required modification of the model and thus, were not te sted. Due to model modification, four additional hypotheses were newly constructe d and the results were all gr oup invariant. Please refer to

PAGE 98

98 Research Questions/Hypotheses section for specific hypotheses; statistical results are presented in Table 9. Comparison with the Existing Literature The results of the MG invarian ce test of the current study were som ewhat inconsistent with prior studies. For instance, multiple hypotheses were constructed based on a premise that depending on how an individual is personally invol ved, the strength of the relationship between one factor and another would var y. Literature partially supported this premise in that when an individual is more involved with a product or situati on, the person is more likely to perceive greater personal relevan ce with the product or situation (Zaichkowsky, 1985). As a result, the person is more likely to engage in a behavior (e.g., purchase of LTM). However, many of the proposed structural relationships we re invariant, and these findings are inconsistent with existing literature. For instance, in the cu rrent study, the structural path (PA IPB) was invariant between HIand LOW-PI groups. Howard and Sh eth (1969) indicated that product involvement would increase differences in the perception of product attributes, product importance, and commitment to brand selection, but inconsistent results were found in the current study. The literature further supported the premise (potential noninvariance of the structural relationships due to varying group memberships) that due to di fferences in the levels of personal involvement, there could be differences in many of the propos ed relationships includi ng attitude, expectancy disconfirmation, satisfaction, and overt behavior More specifically, Celsi and Olson (1988) indicated that higher personal relevance is more likely to have an effect on cognitive behaviors such as attention and comprehension processes, which increases the likelihood of overt behaviors such as shopping and searching. Further, Oliv er and Bearden (1983) studied the role of involvement in satisfaction processes and concluded that involvement is likely to raise expectations prior to product use and will contin ue to have an effect in post-usage evaluation.

PAGE 99

99 Therefore, it is reasonable to assert that level of satisfaction is at least partially determined by ones level of involvement. However, in the current study, the structural paths (SA AB, SA IPB, and SA PA) were invariant between HIand LO W-PI groups. Oliver and Bearden also indicated that involvement positiv ely influenced preand post-usage variables such as attitude, intention, and disconfirmation of expectation and thus, it was e xpected that the structural relationships would vary due to differences in the levels of PI Shaffer and Sherrell (1997) supported this premise, indi cating that no significant relationship was found in a low involvement situation, while expectancy disconfir mation explained the large amount of variance in satisfaction under a high involve ment situation. However, th e hypothesized structural paths (AB IPB and ED SA) were invariant between HIand LOW-PI groups in the current study. Several potential reasons for these inconsistent findings between the curren t study and the existing literature could exist. First, use of a student sample might have influenced the overall results. Scales used in the cu rrent study were designed to ask about LTM in particular; however, if the student sample did not represent the popu lation of LTM consumers, the overall results might have been negatively influenced. S econd, an independent c onstruct (Zaichkowskys product involvement scale) with f our items was used to quantitatively measure PI and used to conduct the measurement test in the current study. However, diffe rent aspects of involvement could exist and thus might have provided different results in the current study. For instance, Laverie and Arnett (2000) measured both situational and enduring involvement and used them to examine the impact on satisfaction. Although the involvement scale in the current study had good reliability and discriminant validity, its co ntent validity might have not been perfectly established. Likewise, although many of these may be mere suppositions, each of these three

PAGE 100

100 aspects or combination of any of them might ha ve generated inconsistent results between the current study and the existing literature. In sum, measurement invariance exists be tween HIand LOW-PI groups, indicating the constructs have been measured equally between the two groups (refer to step 4). Because measurement invariance existed, we were able to conduct structural invariance tests to see which specific parameter did or did not have inva riance between HIand LOW-PI groups. In the current study, structural invariance existed in all structur al paths between HIand LOW-PI groups, indicating differences in the tested pa rameters may not exist between the two groups (refer to step 5). Because an invariance test ha s never appeared in th e literature within the context of sport, it is unfortuna tely not possible to directly co mpare the findings of the current study to the existing studies. This study can be differentiated from previous studies in that the current study merged various theories and conceptualizations into a comprehensive framework. As a result, the proposed model extends previous work by providing a collectivistic vi ew of structural relationships among the factors that lead the c onsumption of LTM. In the current study, although many of the original hypotheses were not supporte d by the statistical resu lts, it is worth noting that the MG invariance test results indicate that the propos ed model is stable. Indeed, establishing equality is a common objectiv e of the MG invariance test. Thus, the current study is unique in that no study exists with in the context of sport that em pirically examined (in)variance of a model across groups. Although further improvement is needed, validation of the scales utilized in the current study provi des practitioners with a reliable and valid me asurement tool that assesses consumers purchase of LTM. Practical im plications of the findings are overviewed in the following sections.

PAGE 101

101 Practical Implications Although many of the discussed theories have been applied to explain a form of sport consumption (frequently game attendance), in orde r to increase applicability, it is crucial to develop a model that incorporates various independent constructs with in the context of LTM consumption. Comprehensiveness of such a mo del would allow scholars and researchers to develop theories that explain various commonly occurring cons umption activities at a domain level. In turn, this research effort will enable retailers to effectively communicate with sport product consumers, which consequently will increase overall sales of LTM. The findings of this study reveal whether e ach of the proposed theoretical constructs contributes to elucidate a specific consumption activity, purchasing LTM. Given the influence of merchandise sales on the overall sport industry, marketers of LTM should search for ways to promote psychosocial constructs including personal values, team identification, attitude toward brand, satisfaction of consumers, perceived attrib utes of products, and pu rchase intention. More specifically, the finding that person al values impact team identif ication and brand attitude is helpful for sport marketers to develop effective marketing strategies. For instance, we now know sport consumers who purchase LTM tend to be patriotic, ambitious, and hedonistic, and these personal values influence higher level of team id entification as well as brand attitude. It is evident that sport industry is being globalized (e.g., Olympic Games, soccer World Cup, World Baseball Classic), and as a result of this, billions of dollars are being spent on LTM within the international market (Foster, Greyser, & Walsh, 2006; Shank, 2009). Thus, utilization of a personal value such as patriotism in developi ng a marketing plan is necessary. Promoting LTM that represent ones own country by emphasizing the value of patriotism is a modern sport business trend.

PAGE 102

102 The finding that team identification impacts br and attitude negatively may provide useful information for sport marketers to select appropr iate brand for the products that represent their team or athletic department. Given that inform ation, it is reasonable to assume that when an individual displays low identification with a team it would be more likely that the person to have stronger attitude toward brand (e.g., NIKE), indicating negative rela tionship between team identification and brand attitude. In this circumst ance, sport marketers will need to emphasize the equity of the brand (strong positive) to prom ote more of product consumption rather than persistently appealing to team identification. In contrast if an individual displays high identification with a team, it would be effective fo r sport marketers to take the advantage of the team identification rather than spending much money to create brand attitude. The finding that perception of product attributes were influenced by fa ctors such as team identification, past expenditure, and satisfaction, and its subsequent impact on the purchase intention of brand is helpful for sport market ers because of the importance of perception of product attributes in purchase decision. More specifically, consumers often consider various product attributes such as price, craftsmanship, aesthe tic color/design, nostalg ia, prestige/status, etc. as important features when making product purchase decision, and thus, sport marketers also often emphasize these features wh en manufacturing products. It w ould be even more effective when sport marketers would know, as the current study identified, what factors influence the formation of ones perception abou t product attributes. It is also worth noting that satisfaction contributed most of the variance in perceived product attributes in the current study. Satisfaction with previous purchases is a powerful impact factor that influences not only consumers perception of product attributes but also brand attitude as well as intention to purchase brand. Therefore, sport marketers will constantly need to monitor the level of fan satisfaction and gather

PAGE 103

103 information about consumers experiences w ith the product, and re flect all the obtained information to meet consumers needs. The fo llowing section discusses several limitations we have identified, and recommendations for each limitation for future study. Limitations and Recommendations for Future Study The current study used a student sample, wh ich m ay have limited its generalizability. Splitting the sample for the purpose of MG invari ance test might have negatively influenced the overall results as well. More specifically, H Iand LOW-PI groups were divided using mean values. However, if the discrepancy was not large enough to distinguish the two groups, it might have influenced the MG invariance test results. In other words, actual differences could have existed among the structural paths but due to insufficient discrepancy between HIand LOW-PI, they were not detected. Thus, it is recommended that a future study increase the sample size, which may widen the gap between th e two groups. Increasing sample size is further rationalized in that due to missing values, only 200 (HI-PI group) and 168 (LOW-PI group) were used for simultaneous SEM analyses. The frequency distribution of the residual covariances should be symmetric because nonsymmetrically distributed residuals in the frequency distribution may signal a poor-fitting model (Tabachnick & Fidell, 2001). However, nonsymmetrical distribution of residuals was found among the results of the current study. To reduce negative influence of these results, multiple fit indices were carefully examined to estimate the model in each step. Eliminating a few cases that contribute large residual values should also be considered. In the SEM using the LOW-PI group (step 3b ), a factor loading of 1.0 was found (PVPA PV). No specific reason could be found that may explain this unrealistic value, and we recognize it as a limitation of the current study.

PAGE 104

104 The respondents were asked to indicate the brand for a piece of LTM that was purchased most recently within the last year. However, ma ny of the respondents identified multiple brands that they purchased in the past (e.g., NIKE, Cham pion, Adidas, Reebok, K-Swiss, Asics, etc.). It is possible that the participants confounded th e brands when responding to the items, thus causing inaccuracies. Future studies should be ca refully designed to reflect this potential limitation. Use of qualitative methods to iden tify additional factors that influence the consumption of LTM is also recommended. Lastly, in the process of model modificati on, three scales, PVCO, AP, and IPP, were eliminated due to lack of either reliability a nd discriminant validity, or combination of the two. Future studies need to include all the eliminated constructs into the m odel, and reevaluate the measurement model. A newly established mode l may provide more insight into how an individual consumer behaves when purchasi ng LTM. The following section summarizes the overall findings and makes a conclusion. Summary and Conclusion Based on values th eory, identity theory, at titude theory, satisfaction theory, and other concepts that influence product consumption, th is study proposed and examined a structural model in an attempt to explain consumption of licensed team merchandise. The empirical findings of the current study suggest ed that consumers intention to purchase of licensed team merchandise is affected by various factors including personal va lues, team identification, brand attitude, past expenditure, pe rceived product attributes, exp ectancy disconfirmation, and satisfaction. Further, MG measurement invariance test results suggested that the same proposed constructs were being measured equally in each group. MG struct ural invariance test results suggested that the structural re lationships among the tested constructs did not vary due to the difference in the level of personal involvement. This finding was inconsistent with the literature

PAGE 105

105 in general. However, it is worth noting that equality was established in the both measurement and structural model, suggesting stability of the proposed model. It is also worth noting that the test of MG invariance of a model is unique because it has not been a common way of model testing within the contex t of sport, making this study more worthwhile. Although there is a need for additional improvement, validation of the scales used in the current study provides researchers with a reliable and valid tool that measures consumers purchase of licensed team merchandise.

PAGE 106

106 APPENDIX A PSYCHOMETRIC PROPERTIES (PILOT TEST RESULTS) Cronbachs alpha, item -to-total correlation, and Cronbachs alpha if item deleted ( N = 80) Item if item deleted Item-to-total correlation Personal values (PV) Conservatism (PVCO) Conventionality is an important value to me (PVCO1) Compliance is an important value to me (PVCO2) Tradition is an important value to me (PVCO3) Conformity is an important value to me (PVCO4) Social norm is an important value to me (PVCO5) Hedonism (PVHE) Self-indulgence is an important value to me (PVHE1) Sensuous gratification is an important value to me (PVHE2) Hedonism is an important value to me (PVHE3) Pleasure is an important value to me (PVHE4) Patriotism (PVPA) Nationalism is an important value to me (PVPA1) Loyalty to country is an important value to me (PVPA2) Patriotism is an important value to me (PVPA3) Devotion to my country is an important value to me (PVPA4) Ambition (PVAM) Ambition is an important value to me (PVAM1) Competition is an important value to me (PVAM2) Aspiring to greatness is an important value to me (PVAM3) Striving to get what I want is an important value to me (PVAM4) Aggressively pursuing my goals is an important value to me (PVAM5) .790 .690 .764 .741 .759 .768 .771 .721 .725 .589 .533 .658 .697 .769 .733 .652 .669 .726 .715 .631 .758 .636 .544 .523 .502 .657 .643 .529 .627 .435 .344 .457 .515 .664 .633 .428 .451 .681 .346 .657 Team identification (TI) I consider myself to be a real fan of the team (TI1) I would experience a loss if I had to stop being a fan of the team (TI2) Being a fan of the team is very important to me (TI3) .842 .809 .823 .713 .679 .688 .788 Attitude toward brand (AB) I like the brand (AB1) I think that the brand is good (AB2) I think that the brand is desirable (AB3) My feelings toward the brand are positive (AB4) .878 .845 .840 .817 .868 .753 .749 .803 .675 Attitude toward product (AP) I like the piece of team licensed merchandise I purchased (AP1) I think that the team licensed merchandise I purchased is good (AP2) I think that the team licensed merchandise I purchased is desirable (AP3) My feelings toward the team licensed merchandise are positive (AP4) .776 .785 .693 .666 .735 .457 .651 .680 .556

PAGE 107

107 Continued. Item if item deleted Item-to-total correlation Intention to purchase (IP) Of product (IPB) In the future, purchasing team licensed merchandise is something I plan to do (IPB1) In the future, I am likely to purchase more of that teams licensed merchandise (IPB2) In the future, I intend to purchase more licensed merchandise representing the team (IPB3) Of brand (IPP) In the future, I intend to purchase more of the same brand of team licensed merchandise (IPP1) In the future, I am likely to purchase more team licensed merchandise of the same brand (IPP2) In the future, I intend to purchase more team licensed merchandise of the same brand (IPP3) .75 .87 .647 .711 .658 .815 .825 .800 .603 .555 .601 .748 .738 .762 Expectancy disconfirmation (ED) The overall quality of the team licensed merchandise was (ED1) The value of the team licensed merchandise was (ED2) The overall performance of the team licensed merchandise was (ED3) The attributes of the team licensed merchandise were (ED4) The benefits of the team licensed merchandise were (ED5) .844 .811 .833 .783 .802 .832 .661 .574 .764 .687 .591 Perceived product attributes (PA) In general, what I get from team licensed merchandise is worth the cost (PA1) All things considered (price, time, and effort), team licensed merchandise is a good buy (PA2) In general, compared to other products, team licensed merchandise is a good value for the money (PA3) In general, team licensed mercha ndise consistently performs better than other products (PA4) In general, team licensed merchandise is aesthetically pleasing (PA5) .737 .752 .668 .636 .743 .645 .327 .570 .647 .382 .616 Satisfaction (SA) I was satisfied with my decision to buy the team licensed merchandise (SA1) I was satisfied with the quality of the team licensed merchandise (SA2) I was satisfied with the value of the team licensed merchandise (SA3) My choice to buy the team licensed merchandise was a wise one (SA4) I think that I did the right thing when I decided to buy the team licensed Merchandise (SA5) .827 .791 .804 .803 .768 .797 .658 .585 .613 .704 .609 Product involvement (PI) To me, my most recently purchased team licensed merchandise is Important (PI1) To me, my most recently purchased team licensed merchandise is relevant (PI2) To me, my most recently purchased team licensed merchandise means a lot to me (PI3) To me, my most recently purchased team licensed merchandise is valuable (PI4) .837 .788 .820 .759 .805 .682 .608 .745 .642

PAGE 108

108 APPENDIX B CORRELATIONS FOR LATENT FACT ORS ( PILOT TEST RESULTS)

PAGE 109

109Pilot test result: Correlati ons for latent factors ( N = 80) 1 2 3 4 5 6 7 8 9 10 11 12 Patriotism (1) 1 Conservatism (2) .315** 1 Ambition (3) .330** .191 1 Hedonism (4) -.045 .453** .297* 1 Team identification (5) .222 -.193 .302* -.244* 1 Brand attitude (6) .199 .203 .215 .152 -.147 1 Product attitude (7) .423** .118 .416** .143 .374** .302* 1 Intention to purchase of product (8) .410** -.076 .423** -.090 .432** .305** .704** 1 Intention to purchase of brand (9) .279* .122 .306** .032 .081 .691** .528** .608** 1 Product attributes (10) .190 .235 .076 .217 .140 .193 .505** .416** .397** 1 Satisfaction (11) .335** .089 .296* .228 .127 .279* .772** .674** .483** .649** 1 Expectancy disconfirmation (12) .158 .313** .132 .183 .062 -.003 .162 -.028 .014 .308** .057 1 Note ** Correlation is significant at the 0.01 level (2-tailed). Correlation is sign ificant at the 0.05 level (2-tailed).

PAGE 110

110 APPENDIX C PERSONAL VALUES ITEMS Personal values item s (Lee & Trail, 2007) Items Scale Strongly Disagree Neutral Strongly Agree Patriotism 1 2 3 4 5 6 7 Nationalism is an important value to me 1 2 3 4 5 6 7 Loyalty to country is an important value to me 1 2 3 4 5 6 7 Patriotism is an important value to me 1 2 3 4 5 6 7 Devotion to my country is an important value to me (new) 1 2 3 4 5 6 7 Ambition Ambition is an important value to me 1 2 3 4 5 6 7 Competition is an important value to me 1 2 3 4 5 6 7 Aspiring to greatness is an important value to me (new) 1 2 3 4 5 6 7 Aggressively pursuing my goals is an important value to me (new) 1 2 3 4 5 6 7 Conservatism Conventionality is an important value to me 1 2 3 4 5 6 7 Tradition is an important value to me 1 2 3 4 5 6 7 Conformity is an important value to me 1 2 3 4 5 6 7 Social norm is an important value to me 1 2 3 4 5 6 7 Hedonism Self-indulgence is an important value to me 1 2 3 4 5 6 7 Sensuous gratification is an im portant value to me 1 2 3 4 5 6 7 Hedonism is an important value to me 1 2 3 4 5 6 7 Pleasure is an important value to me 1 2 3 4 5 6 7 Note Items will be randomly reordered when it is used to collect data.

PAGE 111

111 APPENDIX D TEAM IDENTIFICATION ITEMS Team identification items: Items Scale Strongly Disagree Neutral Strongly Agree I consider myself to be a real fan of the team 1 2 3 4 5 6 7 I would experience a loss if I had to stop being a fan of the team 1 2 3 4 5 6 7 Being a fan of the team is very important to me 1 2 3 4 5 6 7 Note Trail and James (2001) the team identification index (TII)

PAGE 112

112 APPENDIX E ATTITUDE TOWARD BRAND/PR ODUCT ITEMS Attitude toward brand items: Items Scale Strongly Disagree Neutral Strongly Agree I like the brand 1 2 3 4 5 6 7 I think that the brand is good 1 2 3 4 5 6 7 I think that the brand is desirable 1 2 3 4 5 6 7 My feelings toward the brand are positive. 1 2 3 4 5 6 7 Note To measure attitude toward brand, Olivers (1981) attitude toward general product items were modified. Attitude toward product items: Items Scale Strongly Disagree Neutral Strongly Agree I like the piece of team licensed merchandise I purchased 1 2 3 4 5 6 7 I think that the team licensed merchandise I purchased is good 1 2 3 4 5 6 7 I think that the team licensed merchandise I purchased is desirable 1 2 3 4 5 6 7 My feelings toward the team licensed merchandise are positive 1 2 3 4 5 6 7 Note To measure attitude toward brand, Olivers (1981) attitude toward general product items were modified.

PAGE 113

113 APPENDIX F INTENTION TO PURCHASE ITEMS Inten tion to purchase items: Items Scale Strongly Disagree Neutral Strongly Agree In the future, purchasing team licensed merchandise is something I plan to do 1 2 3 4 5 6 7 In the future, I intend to purchase more of the same brand of team licensed merchandise 1 2 3 4 5 6 7 In the future, I am likely to purchase more team licensed merchandise of the same brand 1 2 3 4 5 6 7 In the future, I am likely to purchase more of that teams licensed merchandise 1 2 3 4 5 6 7 In the future, I intend to purchase more team licensed merchandise of the same brand 1 2 3 4 5 6 7 In the future, I intend to purchase more licensed merchandise representing the team 1 2 3 4 5 6 7 Note. Hagger, Chatzisarantis, and Biddles (2001) behavioral intention items were modified. Items #1, 4, and 6 measure intention to purchase of product, whereas items #2, 3, and 5 measure intention to purchase of brand.

PAGE 114

114 APPENDIX G PAST PRODUCT EXPENDITURE ITEMS Past product expenditure item : How much money did you spend over the last ye ar on team licensed merchandise (e.g., sweat shirts, jackets, hats, T-shirts) for yourself fr om that particular brand (e.g., NIKE, Adidas, Reebok)? $______________________________

PAGE 115

115 APPENDIX H EXPECTANCY (DIS)CONFIRMATION ITEMS Expectancy (dis) confirmation items: Items Scale Please indicate the extent to which your expectations for the piece of team licensed merchandise you purchased were met by circling the appropriate number in the scale next to the items below Much Worse than Expected As Expected Much Better than Expected The overall quality of the team licensed merchandise was 1 2 3 4 5 6 7 The overall performance of the team licensed merchandise was 1 2 3 4 5 6 7 The attributes of the team licensed merchandise were 1 2 3 4 5 6 7 The benefits of the team licensed merchandise were 1 2 3 4 5 6 7 Note Reference: modified from Trail, Anders on, & Fink (2005, p. 104). The original items had Cronbachs alpha equals to .91 and AVE value of .67.

PAGE 116

116 APPENDIX I SATISFACTION ITEMS Satisf action items: Items Scale Strongly Disagree Neutral Strongly Agree I was satisfied with my decision to buy the team licensed merchandise 1 2 3 4 5 6 7 I was satisfied with the value of the team licensed merchandise 1 2 3 4 5 6 7 My choice to buy the team licensed merchandise was a wise one 1 2 3 4 5 6 7 I think that I did the righ t thing when I decided to buy the team licensed merchandise 1 2 3 4 5 6 7 Note Items were modified from Oliver (1980).

PAGE 117

117 APPENDIX J PERCEIVED PRODUCT ATTRIBUTES ITEMS Perceived product attributes item s (Netemeyer, Krishnan, Pullig, Wang, Yagci, Dean, Ricks, & Wirth, 2004): Items Scale Strongly Disagree Neutral Strongly Agree In general, what I get from team licensed merchandise is worth the cost (1) 1 2 3 4 5 6 7 All things considered (price, time, and effort), team licensed merchandise is a good buy (2) 1 2 3 4 5 6 7 In general, compared to ot her products, team licensed merchandise is a good value for the money (3) 1 2 3 4 5 6 7 In general, team licensed merchandise consistently performs better than other products (4) 1 2 3 4 5 6 7 In general, compared to ot her products, team licensed merchandise is of very high quality (5) 1 2 3 4 5 6 7 In general, team licensed me rchandise is aesthetically pleasing (new) 1 2 3 4 5 6 7 Note The items were slightly modified from th e original 8-item perc eived quality/perceived value for cost by Netemeyer, Krishnan, Pullig, Wang, Yagci, Dean, Ricks, & Wirth (2004). The original items were reliable and valid (Cronbach s alpha = .87 .96, correlations among the items ranged .41 to .79 indicating discriminant validity, and all items AVE values were greater than .50 indicating construct reliability). The items 1, 2, and 3 represent perceived value for cost, and 4 and 5 represent perceived quality.

PAGE 118

118 APPENDIX K PERSONAL INVOLVEMENT ITEMS Personal inv olvement items: Items Scale Strongly Disagree Neutral Strongly Agree To me, purchasing team licensed merchandise is important 1 2 3 4 5 6 7 To me, purchasing team licensed merchandise is relevant 1 2 3 4 5 6 7 To me, purchasing team licensed merchandise means a lot 1 2 3 4 5 6 7 To me, purchasing team licensed merchandise is essential 1 2 3 4 5 6 7 Note Modified from Zaichkowskys (1994) involvement with product scale (p. 70). The original scale had 10 question items. The original scale had prope rties of the followings : Content validity checked (p. 61), test-retest reliabilities ra nged from .73-.84 (pp. 61-62), and Cronbachs alpha ranged from .91-.96 (pp. 61-62).

PAGE 119

119APPENDIX L PSYCOMETRIC PROPERTIES FROM THE LI TERATRE Reliabilities and validities of th e scales in previous studies Reliability Construct validity Criterion validity Cronbachs alpha Test-retest Construct Convergent Discriminant Concurrent Predictive Values typology scale (Lee & Trail, 2007) .60 ~ .91 NA AVE = 43 ~ .78 Claimed Claimed Claimed NA Team identification index (TII; Trail & James, 2001) .85. (Trail & James, 2001), .88 (Trail, Fink, & Anderson, 2003), .83 (Trail, Anderson, & Fink, 2005) NA AVE = .71 (Trail, Fink, & Anderson, 2003), .62 (Trail, Anderson, & Fink, 2005) NA NA Claimed by comparing correlation coefficients (Trail & James, 2001) NA Attitude items (Oliver, 1981) .85 (Hagger, Chatzisarantis, & Biddle, 2001), .79 (Oliver & Linda, 1981) NA AVE = .72 (Hagger, Chatzisarant is, & Biddle, 2001) NA Claimed (Hagger, Chatzisarantis, & Biddle, 2001) NA NA Intention items (Hagger, Chatzisarantis, & Biddle, 2001) .77 NA AVE = .57 NA Claimed NA NA Expectancy (dis)confirmation items (Trail, Anderson et al., 2005) .79 (Oliver & Linda, 1981), .65 ~ .89 (Oliver, 1993), .76 (Madrigal, 1995), .91 (Trail, Anderson et al., 2005) NA AVE = .67 (Trail, Anderson et al., 2005) NA NA NA NA Note NA: not available.

PAGE 120

120Reliabilities and validities of the scal es in previous studies. Continued. Reliability Construct validity Criterion validity Cronbachs alpha Test-retest Construct Convergent Discriminant Concurrent Predictive Satisfaction items (Oliver, 1980, 1981; Westbrook & Oliver, 1981) .75 ~ .91 (Oliver, 1980; Westbrook & Oliver, 1981), .94 (Oliver & Linda, 1981), .92 ~ .98 (Oliver, 1993), .95 (Madrigal, 1995) NA NA NA NA NA NA Perceived product attributes items (Netemeyer, Krishnan, Pullig, Wang, Yagci, Dean, Ricks, & Wirth, 2004) .87 ~ .96 NA AVE = .66 ~ .79 NA correlations among the items ranged .41 to .79 (recommended to be < .85; Kline, 2005) NA NA Personal involvement items (Zaichkowsky, 1994) .91 ~ 96 .73 ~ .84 NA NA NA NA NA Note NA: not available.

PAGE 121

121 LIST OF REFERENCES Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theo retical analysis and review of e mpirical research. Psychological Bulletin, 84 888-918. Allport, G. W. (1937). Personality: A psyc hological interpretation. New York: Holt. Anderson, J. C., & Gerbing, D. W. (1988). Struct ural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103 411-423. Bagozzi, R. P., & Warshaw, P. R. (1990). Trying to consume. Journal of Consumer Research, 17, 127-140. Ball, A. D., & Tasaki, L. H. (1992). The role and measurement of attachment in consumer behavior. Journal of Consumer Psychology, 1 (2), 155-172. Beatty, S. E., Kahle, L. R., Homer, P. & Misr a, S. (1985). Alternative measurement approaches to consumer values: The List of Values and the Rokeach Value Survey. Psychology & Marketing, 2(Fall), 181-200. Belk, R. W. (1984). Three scales to measure c onstructs related to ma terialism: Reliability, validity, and relationships to measures of happiness. Advances in Consumer Research 11(1), 291-296. Belk, R. W. (1988). Possessi ons and the extended self. Journal of Consumer Research 15(September), 139-168. Bennett, P. D., & Harrell, G. D. (1975). The role of confidence in unde rstanding and predicting buyers attitudes and purchase intentions. Journal of Consumer Research, 2 (September), 110-117. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107 238-246. Bentler, P. M. (2005). EQS 6 structural equa tions program manual. Encino, CA: Multivariate Software. Bloch, P. H., & Richins, M. L. (1983). A theo retical model for the study of product importance perceptions. Journal of Marketing, 47 69-81. Bollen, K. A. (1989). Structural equations with latent variables John Wiley & Sons, Inc. Boninger, D. S., Krosnick, J. A., & Berent, M. K. (1995). The causes of attitude importance: Self-interest, social id entification, and values. Journal of Personality and Social Psychology, 68, 61-80.

PAGE 122

122 Braithwaite, V. A., & Law, H. G. (1985). Struct ure of human values: Tes ting the adequacy of the Rokeach Value Survey. Journal of Personality and Social Psychology, 49 250-263. Broughton, D. (2005, August 29 September 4). Minors score with sales. Sports Business Journal (p. 9), 8(17). Browne, M. W., & Cudeck, R. (1992). Alte rnative Ways of A ssessing Model Fit. Sociological Methods and Research, 21 (2), 230-258. Byrne, B. M. (2006). Structural equation modeling with EQS: Basic concepts, applications, and programming (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Byrne, B. M., & Campbell, T. L. (1999). Cro ss-cultural comparisons and the presumption of equivalent measurement and theoretical structure. Journal of Cross-cultural Psychology, 30(5), 555-574. Byrne, B. M., & Watkins, D. (2003). The i ssue of measurement invariance revisited. Journal of Cross-cultural Psychology, 34 (2), 155-175. Campbell, D. T., & Fiske, D. W. (1959). Convergent and discrimi nant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56 81-105. Caro, L. M., & Garcia, J. A. M. (2006). Cognitiveaffective model of consumer satisfaction: An exploratory study within the fr amework of a sporting event. Journal of Business Research, 6 0, 108-114. Celsi, R. L., & Olson, J. C. (1988). The role of involvement in attention and comprehension processes. Journal of Consumer Research, 15, 210-224. Cheung, G. W., & Rensvold, R. B. (2002). Ev aluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9 233-255. Clarke, K., & Belk, R. W. (1978). The effects of product involvement and task definition on anticipated consumer effort. Advances in Consumer Research, 5 313-318. Cornwell, T. B., & Coote, L. V. (2005). Cor porate sponsorship of a cause: The role of identification in purchase intent. Journal of Business Research, 58 268-276. Cunningham, G. B., & Kwon, H. (2003). The theory of planned behavior and intentions to attend a sport event. Sport Management Review, 6 127-145. Deng, X., Doll, W. J., Hendrickson, A. R., & Scazzero, J. A. (2005). A multi-group analysis of structural invariance: an illustrati on using the technology acceptance model. Information & Management, 42 (5), 745-759.

PAGE 123

123 Dimanche, F., Havitz, M. E., & Howard, D. R. (1991). Testing the involvement profile (IP) scale in the context of selected recreational and touristic activities. Journal of Leisure Research, 51-66. Eagly, A., & Chaiken, S. (1993). The psychology of attitudes Fort Worth: Harcourt Brace Jovanovich College. Engel, J. F., Kollat, D. T., & Blackwell, R. D. (1968). Consumer Behavior. New York: Holt, Rinehart & Winston. Fazio, R. H., Powell, M. C., & Williams, C. J. (1989). The role of attitude accessibility in the attitude-to-behavior process. Journal of Consumer Research 16, 280-289. (B: see marketing class binder) Fazio, R. H., & Zanna, M. P. (1978). On the predicti ve validity of attitudes: The roles of direct experience and confidence. Journal of Personality, 46(2), 228-243. Feather, N. T. (Ed.). (1982). Expectations and actions: Expect ancy-value models in psychology Hillsdale, New Jersey: Lawrence Erlbaum Associates. Feather, N. T. (1982). Human values and the prediction of action: An expectancy-valence analysis. In N. T. Feather (Ed.), Expectations and actions: Ex pectancy-value models in psychology (pp. 263-289). Hillsdale, New Jersey : Lawrence Erlbaum Associates. Feather, N. T. (1992). Values, valences, expectations, and actions. Journal of Social Issues, 48(2), 109-124. Fishbein, M., & Ajzen, I. (1975). Be liefs, attitudes, intentions a nd behavior: An introduction to theory and research. Readi ng, Mass.: Addison-Wesley. Funk, D. C., & James, J. (2001). The psychologi cal continuum model: A conceptual framework for understanding an individuals ps ychological connection to sport. Sport Management Review 4, 119-150. Funk, D.C., & Pastore, D. L. (2000). Equating attitu des to allegiance: The usefulness of selected attitudinal information in segmenting l oyalty to professional sports teams. Sport Marketing Quarterly, 9 (4), 175-184. Funk, D.C., Ridinger, L. L., & Moorman, A. M. (2004). Exploring orig ins of involvement: Understanding the relationship between consumer motives and involvement with professional sport teams. Leisure Sciences, 26 35-61. Gladden, J. M., Irwin, R. L., & Sutton, W. A. (2001). Managing North American Major Professional Sport Teams in the New Mill ennium: A Focus on Building Brand Equity. Journal of Sport Management, 15, 297-317.

PAGE 124

124 Gladden, J. M., & Milne, G. R. (1999). Examining the Importance of Brand Equity in Professional Sport. Sport Marketing Quarterly 8 (1), 21-29. Gladden, J. M., Milne, G. R., & Sutton, W. A. (1998). A conceptual fr amework for assessing brand equity in Division I college athletics. Journal of Sport Management 12, 1-19. Graeff, T. R. (1997). Comprehending product attributes and benefits: The role of product knowledge and Means-End chain inferences. Psychology & Marketing, 14 (2), 163-183. Gutman, J. (1982). A means-end chain model ba sed on consumer categorization processes. Journal of Marketing, 46 60-72. Hagger, M. S., Chatzisarantis, N., & Biddle, S. J. H. (2001). The influence of self-efficacy and past behavior on the physical act ivity intentions of young people. Journal of Sports Sciences, 19 711-725. Hair, J. F., Jr., Anderson, R. E., Ta tham, R. L., & Black, W. C. (1998). Multivariate data analysis (5th ed.). Upper Saddle River, NJ: Prentice-Hall, Inc. Havitz, M. E., & Howard, D. R. (1995). How enduring in enduring involvement? A seasonal examination of three recreational activities. Journal of Consumer Psychology, 4 (3), 255276. Homer, P. M., & Kahle, L. R. (1988). A structur al equation test of the value-attitude-behavior hierarchy. Journal of Personality & Social Psychology, 54 638-646. Houston, M. J., & Rithschild, M. L. (1978). Conceptual and methodological perspectives on involvement. In S. C. Jain (Ed.), Research fr ontiers in marketing: Dialogues and direction (pp. 184-187). Chicago: American Marketing Association. Howard, J. A., & Sheth, J. N. (1969). The theory of buyer behavior New York: John Wiley. Hoyle, R. H. (1995). Structural equation modeling: Concepts, issues, and applications. London: Sage. Hu, L., & Bentler, P. M. (1999). Cutoff criteria fo r fit indexes in covari ance structure analysis: Conventional criteria vers us new alternatives. Structural Equation Modeling 6(1), 1-55. Irwin, R. L., Lachowetz, T., Cornwell, T. B., & Clark, J. (2003) Cause-related sport sponsorship: An assessment of spectator belie fs, attitudes, and behavioral intentions. Sport Marketing Quarterly, 12 (3), 131-139. Iwasaki, Y., & Havitz, M. E. (1998). A path analytic model of the relationships between involvement, psychological commitment and loyalty. Journal of Leisure Research, 30 (2), 256-280.

PAGE 125

125 Jolibert, A., & Baumgartner, G. (1997). Valu es, motivations, and personal goals: Revisited. Psychology & Marketing, 14 (7), 675-688. Jreskog, K., & Srbom, D. (1988). LISREL7: A guide to th e program and applications, Chicago: SPSS. Kahle, L. R. (Ed.). (1983). Social Values and Social Change: Adaptation to Life in America. New York: Praeger. Kahle, L. R, Kambara, K. M., & Rose, G. M. (1996). A functional model of fan attendance motivations for college football. Sport Marketing Quarterly 5(4), 51-60. Kahle, L. R., Beatty, S. E., & Homer, P. (1986). Alternative measur ement approaches to consumer values: The List of Values (L OV) and Values and Life Style (VALS). Journal of Consumer Research 13 (December), 405-409. Kahle, L. R., Duncan, M., Dalakas, V., & Aiken, D. (2001). The social values of fans for mens versus womens university basketball. Sport Marketing Quarterly 10(3), 156-162. Kardes, F. R. (1988). Spontaneous inference processes in advertising: The effects of conclusion omission and involvement on persuasion. Journal of Consumer Research, 15 225-233. Keistetter, D. L., & Kovich, G. M. (1997). An involvement profile of Division I womens basketball spectators. Journal of Sport Management, 11 234-249. Keller, K. L. (1993). Conceptualizing, measuring, and managing customer-based brand equity. Journal of Marketing, 57 1-22. Kim, S-S., Scott, D., & Crompton, J. I. (1997). An exploration of the relationships among social psychological involvement, beha vioral involvement, commitmen t, and future intentions in the contest of birdwatching. Journal of Leisure research, 29 (3), 320-341. Kline, R. B. (2005). Principles and Practice of Structural Equation Modeling New York: The Guilford Press. Kuzma, J. R., Veltri, F. R., Kuzma, A. T., & Miller, J. J. (2003). Ne gative corporate sponsor information: the impact on consumer attitudes and purchase intentions. International Sports Journal, 7(2), 140-147. Kwon, H. H., Trail, G. T. & Anderson, D. F. ( 2005). Are multiple points of attachment necessary in predicting cognitive, affective, conative, or beha vioral loyalty? Sport Management Review, 8, 255 270. Kwon, H. H., Trail, G. & James, J. D. (2007). The mediating role of perceived value: Team identification and purchase intent ion of team-licensed apparel. Journal of Sport Management, 21 540-554.

PAGE 126

126 Lapidus, R. S., & Schibrowsky, J. A. (1996). Do the hot dogs taste better when the home team wins? Journal of Satisfaction, Dissatisfac tion and Complaining Behavior, 9 1-11. Laurent, G., & Kapferer, J-N. (1985). Measuring consumer involvement profiles. Journal of Marketing Research, 22(1), 41-53. Laverie, D. L., & Arnett, D. B. (2000). Factor s affecting fan attendan ce: The influence of identity salience and satisfaction. Journal of Leisure Research, 32 225-246. Lee, D., & Trail, G. T. (2007). Development and validation of the value and goal typology scales. Unpublished research report. University of Florida, Gainesville, Florida. Lee, D., Trail, G. T., Kwon, H. H., & Anderson, D. F. (2006). Psychometric Properties of the MVS, PRS, and PERVAL Scales: Consumer Values versus Perceived Value Unpublished manuscript. Leeuwen, L. V., Quick, S., & Daniel, K. (2002). The Sport Spectator Satisfaction Model: A Conceptual Framework for Understandi ng the Satisfaction of Spectators. Sport Management Review, 5 99-128. Lemon, N. (1973). Attitudes and their measurement New York: John Wiley & Sons, Inc. Lutz, R. J. (1977). An experimental investigat ion of causal relations among cognitions, affect, and behavioral intention. Journal of Consumer Research, 3 (March), 197-208. Lutz, R. J., MacKenzie, S. B., & Belch, G. E. ( 1983). Attitude toward the ad as a mediator of advertising effectiveness: Determinants and consequences. Advances in Consumer Research, 10 (1), 532-539. Lynch, J. G. Jr., Marmorstein, H., & Weigold, M. F. (1988). Choices from sets including remembered brands: Use of recalled at tributes and prior overall evaluations. Journal of Consumer Research, 15 169-184. MacCallum, R. C., Brown, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1 (2), 130-149. MacKenzie, S. B., & Lutz, R. J. (1989). An empi rical examination of the structural antecedents of attitude toward the ad in an advertising pretesting context. Journal of Marketing, 53( April), 48-65. Madrigal, R. (1995). Cognitive and affective determinants of fan satisfaction with sporting event attendance. Journal of Leisure Research, 27, 205-227. Madrigal, R. (2001). Social identity effects in a belief-attitude-intentions hierarchy: implications for corporate sponsorship. Psychology & Marketing, 18 (2), 145-165.

PAGE 127

127 Madrigal. R. (2003). Investing an evolving leisur e experience: Antecedents and consequences of spectator affect during a live sporting event. Journal of Leisure Research 75(1), 23-48. Mahony, D. F., & Howard, D. R. (1998). The impact of attitudes on the behavioral intentions of sport spectators. International Sports Journal, 2 (Winter), 96-110. Mahony, D. F., Madrigal, R., & Howard, D. (200 0). Using the psychological commitment to team (PCT) scale to segment sport consumers based on loyalty. Sport Marketing Quarterly, 9 (1), 15-25. Mahony, D. F., & Moorman, A. M. (1999). The impact of fan attitudes on intentions to watch professional basketball teams on television. Sport Management Review, 2 43-66. Mahony, D. F., & Moorman, A. M. (2000). The relationship between the attitudes of professional sport fans and their inte nsions to watch televised games. Sport Marketing Quarterly, 9 131-139. Matsuoka, H., Chelladurai, P., & Harada, M. (2 003). Direct and indir ect effects of team identification and satisfaction on intention to attend games. Sport Marketing Quarterly, 12(4), 244-253. McIntyre, N. (1989). The personal meaning of participation: Enduring involvement. Journal of Leisure Research, 21 167-179. Mead, G. H. (1934). Mind, self, and society. Chicago: University of Chicago Press. Mitchell, A. A., & Olson, J. C. (1981). Are produ ct attribute beliefs th e only mediator of advertising effects on brand attitude? Journal of Marketing Research, 18 318-332. Muthn, B., & Kaplan, D. (1985). A comparison of methodologies for the factor analysis of nonnormal Likert variables. British Journal of Mathematical and Statistical Psychology, 38 171-189. Netemeyer, R. G., Krishnan, B., Pullig, C., Wang, G ., Yagci, M., Dean, D., Ricks, J., & Wirth, F. (2004). Developing and validating measures of facets of customer-based brand equity. Journal of Business Research, 57 209-224. Oliver, R. L. (1980). A cognitive model of the antecedents and consequences of satisfaction decisions. Journal of Marketing Research 17(November), 460-469. Oliver, R. L. (1981). Measuremen t and evaluation of satisfaction processes in retail settings. Journal of Retailing 57(3), 25-48. Oliver, R. L. (1989). Processing of the satisfaction response in consumption: A suggested framework and research propositions Journal of Consumer Sa tisfaction, Dissatisfaction and Complaining Behavior, 2, 1-16.

PAGE 128

128 Oliver, R. L. (1993). Cognitive, affective, and attribute bases of the satisfaction response. Journal of Consumer Research, 20, 418-430. Oliver, R. L. (1997a). Loyalty and profit: Long-term effects of satisfaction. In R. L. Oliver (Ed.), Satisfaction: A behavioral perspective on the consumer (pp. 65-92). New York: McGraw-Hill Companies, Inc. Oliver, R. L. (1997b). Satisfaction: A behavioral perspective on the consumer. New York: McGraw-Hill Companies, Inc. Oliver, R. L. (1999). Whence consumer loyalty? [Special issue]. Journal of Marketing, 63 3344. Oliver, R. L., & Bearden, W. O. (1983). The role of involveme nt in satisfaction processes. Advances in Consumer Research, 10 (1), 250-255. Oliver, R. L., & Linda, G. (1981). Effect of satisfaction and its antecedents on consumer preference and intention. Advances in Consumer Research, 8, 88-93. Park, C. W., & McClung, G. W. (1986). The eff ect of TV program involvement on involvement with commercials. Advances in Consumer Research, 13 544-548. Park, C. W., & Young, S. M. (1986). The impact of involvement and the role of music in advertisements on brand attitude formation. Journal of Marketing Research, 23(February), 11-24. Petty, R. E., & Cacioppo, J. T. (1981). Attitudes and persuasion: Classic and contemporary approaches. Dubuque, IA: William C. Brown., Pe tty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and peripheral routes to advertising effect iveness: The moderating role of involvement. Journal of Consumer Research, 10 135-146. Raykov, T., & Marcoulides, G. A. (2000). A firs t course in structural equation modeling. Mahwah, NJ: Lawrence Erlbaum Associates. Rahim, M. A., & Magner, N. R. (1996). Confirmato ry factor analysis of the bases of leader power: First-order factor model a nd its invariance across groups. Multivariate Behavioral Research, 31 495-516. Richins, M. L. (1994a). Valuing things: The public and private meanings of possessions. Journal of Consumer Research 21(December), 504-521. Richins, M. L. (1994b). Special possessions and the expression of material values. Journal of Consumer Research, 21(December), 522-533.

PAGE 129

129 Richins, M. L. Bloch, P. H., & McQuarrie, E. F. (1992). How enduring and situational involvement combine to create involvement responses. Journal of Consumer Psychology, 1(2), 143-153. Richins, M. L., & Dawson, S. (1992). A consumer values orientation for materialism and its measurement: Scale development and validation. Journal of Consumer Research 19(3), 303-316. Rokeach, M. (1968). The Role of Values in Public Opinion Research. Public Opinion Quarterly, 32, 547-549. Rokeach, M. (1973a). The Nature of Human Values New York: The Free Press. Rokeach, M. (1973b). Understanding Human Values : Individual and Societal New York: The Free Press. Rokeach, M., & Ball-Rokeach, S. J. (1989). Stability and change in American value priorities, 1968-1981. American Psychologist, 44 (5), 775-784. Rothschild, M. L. (1984). Perspectives on involve ment: Current problems and future directions. Advances in Consumer Research, 11 216-217. Roy, D. P., & Graeff, T. R. (2003). Consumer at titudes toward cause-relate d marketing activities in professional sports. Sport Marketing Quarterly, 12 (3), 163-172. Schuett, M. A. (1993). Refining meas ures of recreation involvement. Leisure Sciences, 15 205216. Schwartz, S. H. (1992). Universals in the conten t and structure of values: Theoretical advances and empirical tests in 20 countries. Advances in Experimental Social Psychology, 25, 165. Schwartz, S. H. (1994). Are there universal aspe cts in the structure a nd contents of human values? Journal of Social Issues, 50 (4) 19-45. Schwartz, S. H. (1996). Value priorities and beha vior: Applying a theory of integrated value systems. In Seligman, C., Olson, J. M., & Zanna, M. P (Eds.), The psychology of values: The Ontario Symposium (pp. 1-24). NJ: Lawrence Erlbaum Associates. Schwartz, S. H., & Bilsky, W. (1987). Toward a universal psychological structure of human values. Journal of Personality and Social Psychology, 53 (3), 550-562. Schwartz, S. H., & Bilsk y, W. (1990). Toward a theory of the universal content and structure of values: Extensions and cr oss-cultural replications. Journal of Personality and Social Psychology, 58(5), 878-891.

PAGE 130

130 Schwartz, S. H., & Huismans, S. (1995). Value priorities and religios ity in four western religions. Social Psychology Quarterly, 58 (2), 88-107. Schwartz, S. H., & Sagiv, L. (1995). Identifying cultu re-specifics in the co ntent and structure of values. Journal of Cross-Cultural Psychology, 26 (1), 92-116. Serpe, R. T. (1987). Stability and change in self: A structural sym bolic interactionist explanation. Social Psychology Quarterly, 50, 44-55. Shaffer, T. R., & Sherrell, D. L. (1997). Cons umer satisfaction with health-care services: The influence of involvement. Psychology & Marketing, 14 (3), 261-285. Sherif, M., & Cantril, H. (1947). The Psychology of Ego-Involvement New York: John Wiley & Sons, Inc. Sheth, J. N. (1973). A model of industrial buyer behavior. Journal of Marketing, 37, 50-56. Sheth, J. N., Newman, B, I., & Gross, B. L. (1991). Why we buy what we buy: A theory of consumption values. Journal of Business Research, 22 159-170. Smith, R. E., & Swinyard, W. R. (1983). Attit ude-behavior consistency: The impact of product trial versus advertising. Journal of Marketing Research, 20 (August), 257-267. Smith, G. J., Patterson, B., Williams, T., & Hogg, J. (1981). A profile of the deeply committed male sports fan. Arena Review 5(2), 26-44. Stryker, S. (1968). Identity salie nce and role performance: the re levance of symbolic interaction theory for family research. Journal of Marriage and the Family, 30, 558-564. Stryker, S., & Burke, P. J. (2000). The Past, Present, and Future of an Identity Theory. Social Psychology Quarterly 63 (4), 284-297. Stryker, S., & Serpe, R. T. (1994). Identity salience and psychological centrality: Equivalent, overlapping, or complementary concepts? Social Psychology Quarterly 57 16-35. Sutton, W. A., McDonald, M. A., Milne, G., & Cimperman, J. (1997). Creating and fostering fan identification in professional sports. Sports Marketing Quarterly 6, 15-22. Tajfel, H. & Turner, J. C. (1979). An integrative th eory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The Social Psychology of Intergroup Relations Monterey, CA: Brooks-Cole. Tracey, T. J. G., Watanabe, N., & Schneider, P. L. (1997). Structural invariance of vocational interests across Japanese and American cultures. Journal of Counseling Psychology, 44, 346-354.

PAGE 131

131 Trail, G. T., Anderson, D. F., & Fink, J. S. (2005) Consumer satisfaction and identity theory: A model of sport specta tor conative loyalty. Sport Marketing Quarterly, 14 98-111. Trail, G. T., Anderson, D. F., & Fink, J. S. (2000). A theoretical model of sport spectator consumption behavior. International Journal of Sport Management, 1, 154-180. Trail, G. T., Anderson, D. F., & Lee, D. (2006) Determinants of attendance: The predictive value of team identification, past attendance, and attendan ce intentions. Paper presented at the meeting of the Sport Ma rketing Association, Denver, CO. Trail, G. T., & Chelladurai, P. (2002). Perceptions of intercollegiate athlet ic goals and processes: The influence of personal values. Journal of Sport Management, 16 289-310. Trail, G. T., Fink, J. S., & Anderson, D. F. (2003). Sport spectator consumption behavior. Sport Marketing Quarterly, 12 8-17. Trail, G. T. & James, J. D. (2001). The motiva tion scale for sport consumption: Assessment of the scales psychometric properties. Journal of Sport Behavior, 24 (1), 108-127. Trail, G. T., Robinson, M. J., Dick, R. J., & G illentine, A. J. (2003). Motives and points of attachment: Fans versus spectator s in intercollegiate athletics. Sport Marketing Quarterly, 12(4), 217-227. Vinson, D. E., Scott, J. E., & Lamont, L. M. (19 77). The role of personal values in marketing and consumer behavior. Journal of Marketing 41 (April), 44-50. Wakefield, K. L., & Blodgett, J. G. (1994). The im portance of servicescapes in leisure service settings. Journal of Services Marketing, 8 (3), 66-76. Wann, D. L. (2002). Preliminary Validation of a Measure for Assessing Iden tification as a Sport Fan: The Sport Fandom Questionnaire. International Journal of Sport Management 3, 103-115. Wann, D. L., & Branscombe, N. R. (1993). Sports fans: Measuring degree of identification with their team. International Journal of Sport Psychology, 24 1-17. Wann, D. L., & Branscombe, N. R. (1995). Influe nce of identification wi th a sports team on objective knowledge and subjective beliefs. International Journal of Sport Psychology 26(4), 551-567. Wann, D. L. & Dolan, T. J. (1994a). Attributi ons of highly identified sports spectators. The Journal of Social Psychology 134(6), 783-792. Wann, D. L. & Dolan, T. J. (1994b). Influence of spectators identification on evaluation of the past, present, and future performance of a sports team. Perceptual and Motor Skills 78 547-552.

PAGE 132

132 Wann, D. L., Ensor, C. L., & Bilyeu, J. K. (2001) Intrinsic and extrinsic motives for originally following a sport team and team Identification. Perceptual and Motor Skills 93, 451-454. Wann, D. L., & Robinson III, T. N. (2002). The relationship between sport team identification and integration into and pe rceptions of a university. International Sports Journal, 6 (1), 36-44. Westbrook, R. A., & Oliver, R. L. (1981). Developing better measures of consumer satisfaction: Some preliminary results. Advances in Consumer Research, 8, 94-99. Wicker, A. W. (1969). Attitudes versus actions: Th e relationship of verbal and overt behavioral responses to attitude objects. Journal of Social Issues, 25 41-78. Zaichkowsky, J. L. (1985). Measuring the involvement construct. Journal of Consumer Research, 12 341-352. Zaichkowsky, J. L. (1986). Conceptualizing involvement. Journal of Advertising, 15 4-14. Zaichkowsky, J. L. (1994). The personal involv ement inventory: Reduction, revision, and application to advertising. Journal of Advertising, 23(4) 59-70. Zanna, M. P., & Rempel, J. K. (1988). Attit udes: A new look at an old concept. In The Social Psychology of Knowledge ed. D. Bar-Tal, A. W. Kruglanski, pp. 315-334. New York: Cambridge University. Zhang, J. J., Pease, D. G., & Smith, D. W. ( 1998). Relationship between broadcasting media and minor league hockey game attendance. Journal of Sport Management, 12 103-122. Zhang, J. J., Smith, D. W., Pease, D. G., & La m, E. T. C. (1998). Dimensions of spectator satisfaction toward support programs of professional hockey games. International Sports Journal, 2 1-17. Zhang, J. J., Smith, D. W., Pease, D. G., & Ma har, M. T. (1996). Spectator knowledge of hockey as a significant predicto r of game attendance. Sport Marketing Quarterly, 5(3), 41-48. Zhang, Z., Won, D., & Pastore, D. L. (2005). The effects of attitudes to ward commercialization on college students purchasing inte ntions of sponsors products. Sport Marketing Quarterly, 14 177-187. Zillman, D., Bryant, J., & Saplosky, B. J. (1989). Enjoyment from sport spectatorship. In J. D. Goldstein (Ed.), Sports, games, and play : Social and psychological viewpoints (2nd ed.) (pp. 241-278). Hillsdale, NJ: Lawrence Erlbaum Associates.

PAGE 133

133 BIOGRAPHICAL SKETCH Donghun Lee, who is also called by others as Don, was born on Septem ber 23, 1973 in Seoul, South Korea. He earned his B.S. in physical education and his M.A. in sport management from the KeiMyung University, Daegu, Korea, in 1999 and the Ohio State University, Columbus, Ohio, in 2003, respectively. Don accepted an assistant professor position from the business department in the college of Mount St. JosephCincinnati, Ohio.