同儕認同、線上品牌社群認同與品牌忠誠度的關係：以「品牌─社群融合度」為干擾變數

全文

(1)國立臺灣師範大學管理學院管理研究所碩士論文 Graduate Institute of Management College of Management National Taiwan Normal University Master Thesis. Peer Identification, Online Brand Community Identification and Brand Loyalty: Brand-Community Identity Fusion as a Moderator. Jui-Che Chang. Advisor： Shih-Ju Wang Ph.D.. July, 2013.

(2) Table of Content Table of Content ..................................................................................................................................................................i List of Figures .................................................................................................................................................................... iii List of Tables...................................................................................................................................................................... iv Chapter 1：Introduction ............................................................................................................................................... 1 1.1 Research Background ......................................................................................................................................... 1 1.2 Research Question ............................................................................................................................................... 6 Chapter 2：Theoretical Background and Hypotheses ..................................................................................... 7 2.1 Brand Community Identification ................................................................................................................... 7 2.1.1 Community....................................................................................................................................................... 7 2.1.2 Online Brand Community ........................................................................................................................... 8 2.1.3 Brand Community Identification .......................................................................................................... 12 2.2 Antecedents of Brand Community Identification ................................................................................ 14 2.2.1 Community Support .................................................................................................................................. 15 2.2.2 Community Value Congruence .............................................................................................................. 16 2.2.3 Community Affirmation ........................................................................................................................... 16 2.3 Peer Identification ............................................................................................................................................ 17 2.3.1 Peer Identity Similarity ............................................................................................................................ 18 2.3.2 Peer Identity Distinctiveness .................................................................................................................. 19 2.3.3 Peer Identity Prestige ............................................................................................................................... 20 2.3.4 Peer Identification to Brand Community Identification .............................................................. 20 2.4 Brand Loyalty...................................................................................................................................................... 21 2.5 Moderating role of Identity Fusion ............................................................................................................ 23 Chapter 3：Methodology ........................................................................................................................................... 25 3.1 Research Model .................................................................................................................................................. 25 3.2 Sampling and Data Collection Procedures .............................................................................................. 27 i.

(3) 3.3 Questionnaire Design ...................................................................................................................................... 29 3.4 Measures ............................................................................................................................................................... 30 Chapter 4: Results ......................................................................................................................................................... 34 4.1 Descriptive Statistics........................................................................................................................................ 34 4.1.1 Sample Characteristics ............................................................................................................................. 34 4.1.2 User Behavior Analysis ............................................................................................................................. 36 4.2 Factor Analysis ................................................................................................................................................... 42 4.3 Structural Equation Model ............................................................................................................................ 47 4.3.1 Test of Goodness-of-fit .............................................................................................................................. 47 4.3.2 Parameter estimation............................................................................................................................... 48 4.4 Hypothesis Testing ........................................................................................................................................... 52 Chapter 5: Discussion .................................................................................................................................................. 56 5.1 Research Contributions .................................................................................................................................. 56 5.1.1 Peer identification lead to brand community identification ...................................................... 56 5.1.2 Peer identification lead to brand loyalty, rather than brand community identification . 56 5.1.3 Community support facilitates the forming of peer identification........................................... 57 5.1.4 The brand-community identity fusion strengthen the positive effect member similarity on peer identification .......................................................................................................................................... 57 5.2 Managerial Implications ................................................................................................................................. 58 5.3 Limitations and Directions for Further Research ................................................................................ 59 References ........................................................................................................................................................................ 61 Appendix.............................................................................................................................................................................67. ii.

(4) List of Figures Figure 1.1 Growth of global internet users ............................................................................................................ 2 Figure 1.2 Internet users per 100 inhabitants ..................................................................................................... 2 Figure 1.3 Online activities people do in an hour ............................................................................................... 3 Figure 1.4 Popular internet activities changing over time .............................................................................. 4 Figure 1.5 The reasons for joining brand communities ................................................................................... 5 Figure 2.1 Bhattacharya and Sen’s framework of organizational identification................................. 18 Figure 3.1 Research model ........................................................................................................................................ 25 Figure 3.2 Types of issues discussed through communities ....................................................................... 27 Figure 4.1 Statistical result of survey respondents’ possession of automobile ................................... 37 Figure 4.2 Statistical result of survey respondents’ most-used camera ................................................. 37 Figure 4.3 Statistical result of survey respondents’ length of car usage ................................................ 38 Figure 4.4 Statistical result of survey respondents’ membership of automobile brand communities .................................................................................................................................................................... 38 Figure 4.5 Statistical result of survey respondents’ most-visited automobile brand communities ............................................................................................................................................................................................... 39 Figure 4.6 Path diagram.............................................................................................................................................. 48 Figure 4.7 Parameter estimates for conceptual model .................................................................................. 55. iii.

(5) List of Tables Table 2.1 Four elements definite community ....................................................................................................... 8 Table 3.1 Hypotheses ................................................................................................................................................... 26 Table 3.2 Brand communities as subjects of study ......................................................................................... 28 Table 3.3 Operational definitions and measurement items ........................................................................ 30 Table 4.1 Descriptive statistics of respondents’ characteristics (N=280) ............................................. 35 Table 4.2 Analysis of user behavior (N=280) .................................................................................................... 40 Table 4.3 Factor analysis results ............................................................................................................................. 44 Table 4.4 Correlations among constructs, convergent validity and discriminant validity ............. 46 Table 4.5 Goodness-of-fit statistics for conceptual model ........................................................................... 47 Table 4.6 Standardized estimates on latent endogenous variables and observed variables......... 49 Table 4.7 Decomposition of standardized effects for conceptual model (γ value) ............................. 50 Table 4.8 Decomposition of standardized effects for conceptual model (β value) ............................ 50 Table 4.9 Moderating effect of brand-community identity fusion ............................................................ 51 Table 4.10 Hypotheses testing ................................................................................................................................. 54. iv.

(6) Abstract: Building an online brand community is an important mission for business and companies who want to obtain the marketing benefit, but few of them succeed. This research examines the relationship between the identification (member-to-community and member-to-member) in online brand community and brand loyalty. In addition, we also examine the effect that brand-community identity fusion brought to online brand communities owing to the diversified non-brand related content of them. At last, we propose several advice for practitioners to build an online brand community right.. Keyword: Brand Community, Identification, Peer Identification, Brand Loyalty, Identity Fusion. Chapter 1：Introduction. 1.1 Research Background Internet nowadays becomes a media for information exchange and social interaction and bring a radical change to human’s daily life. Tremendous growth in internet and smartphone usage worldwide these years make the global internet population grow by 1.1 billion from 2007 to 2012 (seeing figure 1.1), and the increase is more obvious in developed countries rather than developing country (seeing figure 1.2).. 1.

(7) Source: Internet World Stats (2012) Figure 1.1 Growth of global internet users. Source: International Telecommunications Union(2012) Figure 1.2 Internet users per 100 inhabitants. 2.

(8) People's online behavior is changing day by day. In August 2011, the UK internet population spent 3.4 billion hours online, figure 1.3 shows the portions of time if we distilled all usage into a single hour. It indicated that they spent most of time in social media, accounted for 14 minutes of an hour. Besides, according to a February 2012 survey examined by Pew Internet and America Life Project shows that 66% of online adults use social network site on the web, up from 11% in 2005(seeing figure 1.4). The social media change the communication and linkage of inter-personal by the features such like producing and delivering by users, more and more channels to communicate, information explosion and highly interactive.. Source: Experian Hitwise (2011) Figure 1.3 Online activities people do in an hour. 3.

(9) Source: The Pew Research Center's Internet & America Life Project (2002-2012) Figure 1.4 Popular internet activities changing over time. More and more company saw the great potential in social media and try to figure out how to connect and interact with customers at any time. They first establish official brand/company website in order to maintain the business model of O2O (online to offline) and gain their benefit. The official website seems to be the best media for communication between company and consumers; however, according to the global survey from University of MaCann (2011) of these four years, the population of consumers visited the official website decreased from 85% to 72%, while 67% in Taiwan.. Although the decline of visiting the official website, the research found that more and more consumer joined particular brand communities or fan club for acquiring information. UM (2010) survey found that the participation rate of specific brand communities and fan club in Taiwan is 51%. Accordingly, marketers in industries are 4.

(10) busy trying to build communities around their own brands. The timing is right because people are hungry for a sense of connection in today’s turbulent world and companies need to do more about their products by new ways. Unfortunately, although many firms seek a strong community to reach marketing efficiency and increase the customer loyalty, few understand what it takes to attain such benefits.. For customers, there are many reasons to join a brand community. A survey from InsightXplorer marketing research (seeing figure 1.5) shows that: After joining the brand community, 72% of customers enhance their sense of belonging to the brand, 71% of them believe they are more likely to purchase the product of the brand, 66% of them feel more loyal to the brand, as well as 63% of them think they would recommend others to join. It is speculated that consumers increase the purchase intention and brand identification after joining brand communities.. Source: InsightXplorer(2010/09) Figure 1.5 The reasons for joining brand communities 5.

(11) Building brand community is a challenging task for every company and there are some research proposed that enhancing community identification of customers will increase the marketing effect such as brand loyalty and prestige, making customers spend on the product and service. Because in the online brand community, members build relationship with not only the community but the other members, this study takes brand communities as samples and divides brand community identification into two parts—to the community and to the members within. Besides, the interaction between members make brand communities more diversity and plentiful, so we examine its' effect to the degree of fusion between the brand and the community. We also observe the impact identification on brand loyalty in such communities.. 1.2 Research Question This paper uses two point of view to explore the identification and its' antecedents in the brand community. Moreover, we also explore the impact of the identification to the brand loyalty.. 1.. What antecedents lead to brand community identification and peer identification?. 2.. Would peer identification lead to brand community identification within an online brand community?. 3.. Would peer identification and brand community identification lead to greater brand loyalty?. 4.. Could brand-community identity fusion moderate the influence member similarity have on peer identification in the context of more diversified community?. 6.

(12) Chapter 2：Theoretical Background and Hypotheses. 2.1 Brand Community Identification. 2.1.1 Community Rothaermel and Sugiyama (2001) suggested that individuals come together as a group based on an obligation to one another or come together as a group to be one in purpose are community (Rothaermel and Sugiyama, 2001). Community phenomenon is a core areas of anthropological research, and many studies have previously proposed almost one hundred definitions of "community". Although there is no definite theory of community, but some scholars still attempt to integrate the wide variety of definition in previous studies. Three elements defining communities identified by Karp, Stone and Yoels (1977) : (1) sustained social interaction, (2) shared attributes and values, and (3) a delineated geographical space (Karp, Stone and Yoels, 1977). McMillan and Chavis(1986) examined a definition of community, including four elements: membership, influence, integration and fulfillment of needs and shared emotional connection (McMillan and Chavis, 1986).. 7.

(13) Table 2.1 Four elements definite community Element Membership. Explanation To feel a sense of belonging or to share a sense of personal relatedness.. Influence. One is mattering to make a difference to a group and the group is important for its members. Integration and. Members can meet their needs by receiving. Fulfillment of needs. resources through the membership in a group.. Shared Emotional. Members have the commitment and belief to share. Connection. the common places and time together, and much the same experiences.. Source: McMillan and Chavis(1986). Two types of communities defined by Gusfield: One is the geographic community like an area, town or district, another is a relational community regarding human relationships like hobby clubs, religious groups and fan clubs (Gusfield, 1978). 2.1.2 Online Brand Community The research of community became broader with the trend of modernity, market capitalism, and consumer culture (Muniz and O’Guinn, 2001). Mechanical advancement as well as scientific and technological progress driven by modernization make community more than a place. Muniz and O’Guinn (2001) emphasized that communities are not restricted by geography any more. It was described as a place at the first, just like rural. Nevertheless, the thought of community breakthrough those limitation and ran over into a much broader field of meaning. Thanks to new communication technologies’ ability, geographically dispersed individuals are brought together with a 8.

(14) commonality of purpose and identity (Muniz and O’Guinn, 2001).. In the consumer behavior area, a periodic community in a farmer’s market has been recorded by McGrath, Sherry, and Heisley (1993). They argued that the members in consumption communities considered some consumer experiences and behaviors as contents and issues in the community, and deemed it as a part of the traditions of the community. For instance, collective purchase (McGrath, Sherry, and Heisley, 1993).. Schouten and McAlexander (1995) proposed the concept of "subculture". The Harley Davidson riders share the connection between each other, it play an important role of making them understand more about the brand. Further, the understanding become a real life style they called subculture (Schouten and McAlexander, 1995). Consumers have shared common interest, values and behaviors in sub groups where they engaged in. As a consequence, these groups bring about a high degree of commitment to a particular product or brand and we can find their subculture in the products they deal with. Their own social structure and core values involve to all facets of their lives. The members' commitment and the product or activities maintain their subculture and bring them together (Schouten and McAlexander, 1995).. Based on Schouten and McAlexander(1995), the concept of a community built specifically around brands has been introduced by Muniz and O’Guinn (2001). A brand community is a bound community that is specialized and non-geographically, consisted with admirers of a brand and based on a set of social relationships(Muniz and O’Guinn 2001). It's possible to form a community like this around any brand, especially a brand which has a strong image, a abounding and lengthy history. Furthermore, threatening competition. Because of the omnipresent essence of brands, this kind of communities 9.

(15) may go beyond geography and may include numerous consumer members. The commitment to both the brand and the group make these social groups quite stable. They would be commercial and provided with a mass-media sensibility (Muniz and O’Guinn 2001).. After observed the communities of computers and cars, three core components of brand community identified by Muniz and O’Guinn (2001) as following: (1) Consciousness of Kind: Members in brand communities feel that they have links with the brand. However, they think that the connection among other members is more stronger. Members think that they somehow know each other even though they have never met. As Cova's (1997) argument that “the link is more important than the thing.” Consciousness of Kind is also represents the intrinsic link between members and drive them distinguish themselves from users of other brands. Below are two elements compose Consciousness of Kind: Legitimacy: Legitimacy is a process to differentiate whether a person as a true member of the community and those who are not. Brand communities do not reject membership as a open social organizations, but they have status hierarchies just like most communities. And that, distinguish between true believer and a passing-fad members is a common issue for the members in the brand community. Oppositional Brand Loyalty: It's also a social process to sustain consciousness of kind. Members in brand community place importance on social experiences and gain understanding of the meaning of the brand through opposition to competing brands. (2) Rituals and Traditions: 10.

(16) Rituals and traditions indicate important social processes in brand community which go around the shared consumption experience with the brand. These brand community rituals and traditions created in order to maintain the culture of the community. Celebrating the History of the Brand: It's important for brand community to transmit the history of brand to members, for the purpose of strengthening the value of community. It's also can encourage the similarity between members. Sharing Brand Stories: Storytelling facilitate to maintain a community. An organization use legend to create different impressions and strengthen the consciousness of kind. Members in the community feel safety after they recognize that there are many like-minded others. Text and symbols are vital elements representing the culture of a group (Gusfield, 1978). The product and logo of the brand , not only contemporary and classic, but also images and text from advertisements, are text and symbols in brand community (Muniz and O’Guinn 2001).. (3) Moral Responsibility: Members have a sense of duty to the entire community, and to every single members of the community. Sense of moral responsibility push ahead the community cohesive and bring collective actions. Two important errands to execute in brand community: Integrating and Retaining Members: To ensure the sustainable operation and development of the community, keep old members stay and arrange for making new members fit in to the community is very important. Through integrating and retaining Members, loyalty to the community and the brand will be perpetuating. 11.

(17) Assisting in the Use of the Brand: Members call for or help other members in the consumption of the brand. They do this without consideration and originate in the sense of responsibility.. The components above-mentioned are proposed by Muniz and O'Guinn(2001), and they defined brand community as a community established at the base of "product" or "service" of a brand. It has not only the three core component that possessed by other general community, but also a social structure of relationship formed by a group of people who advocate the brand, which is specific and non-geographic restriction.. Therefore, the members in brand community can persistently and widely effect the ideas and acts of the brand community (Muniz and O'Guinn, 2001). Moreover, spreading knowledge (Brown et al, 2003), knowing other customers' evaluation of products, increasing opportunities for customers to participate in activities and cooperating with highly loyal customers (Franke and Shah, 2003). 2.1.3 Brand Community Identification The process that individuals categorize themselves into a particular social group is called identification, and it helps themselves recognize that they are special and distinct from others (Bhattacharya and Sen, 2003).. Relate to intergroup relations, group processes, and the social self, social identity theory is a social psychological theory (Hogg et al. 1995). An individual is psychologically aware of a linkage between himself or herself and the group, and sharing the experience of the group through the formation of a social identity. Social identification helps people to answer the question, “Who am I?” (Ashforth and Mael 12.

(18) 1989).. Organizational identification, a specific form of social identification that defined as an individual's perception of belongingness to an organization (Mael and Ashforth 1992). It's also referred to as a critical component that binding individuals to an organization. Mael and Ashforth (1992) have demonstrated that a persons' identification to an organization can not only cause increased support of the organization, but participation in it. Organizational identification lead to increased intention of purchasing products or brands (Aaker 1999) and raise the receptivity and effectiveness of advertising (Wheeler et al. 2005); and further, it reduce the turnover rate of employee (O’Reilly and Chatman 1986), increase customer citizenship behaviors and market share (Ahearne et al. 2005). Seeing that organizational identification leads to positive firm outcomes, constituents' strong identification with the organization become a marketer's goal (Fombelle, Paul W.; Jarvis, Cheryl Burke; Ward, James,. 2012). In other words, Due to much benefit come from organizational identification, the researchers figured out factors that can increase an individual’s organizational identification, such like organizational distinctiveness, organizational prestige, tenure and satisfaction in the organization, and the visibility of the membership (Algesheimer et al. 2005; Bhattacharya et al. 1995). Marketers want to encourage constituents to strongly identify with their organizations, because organizational identification contributes to positive firm outcomes.. Algesheimer, Dholakia and Herrmann (2005) describe "brand community identification" as the strength of the relationship between consumers and the community. This relationship between consumers and the community affects the identification which can bring about engagement and perceptions of normative pressure 13.

(19) and resistance. Finally, lead to various community- and brand-related behaviors.. Consumers consider himself or herself to be a member who "belonging" to the brand community, which is characterized as "brand community identification". This kind of identification is a shared or collective identity and make an individual unique and separate. Algesheimer et al. (2005) defined social identity as the cognitive and affective component (Bagozzi and Bergami, 2000) involved in a valued group such like brand community. With regard to the cognitive component, the brand community identification is related to categorization processes which is a self-awareness of membership formulated and maintains by consumers in the community. Such like "I see myself as part of the community", presents the similarities with other members and the dissimilarities with nonmembers more clearly. The concept is just like the consciousness-of-kind aspect of brand communities (Muniz and O'Guinn 2001). The identification regarding the affective component implies a sense of emotional involvement with the group, which is described as an "affective commitment" to the group (Ellemers, Kortekaas, and Ouwerkerk 1999). The researchers of brand community characterized it as "kinship between members" (McAlexander, Schouten, and Koenig 2002). In short, identification indicates that the consumer agrees with the community's norms, traditions, rituals, and objectives (Bhattacharya, Rao, and Glynn 1995).. 2.2 Antecedents of Brand Community Identification However, although there are many research shows that increasing members’ identification with an organization leads to a lots of positive effects, few have sought to figure out the reason why organizational identification occur. 14.

(20) Fombelle, Jarvis, Ward and Ostrom (2012) synthesized previous research and shown that organizational identification is positively affected by: perceived support (McAlexander et al. 2002), value congruence (Bhattacharya and Sen 2003), and identity affirmation (Drigotas et al. 1999). 2.2.1 Community Support Members feel that they were supported in an organization when their contribution is respected and their existence is cared about by the organization. (Eisenberger et al. 1986). Perceived organizational support represents a psychological contract between members and the organization (Kraimer and Wayne, 2004). When the psychological contract satisfied by the organization, members feel that the organization respect for their contributions (Eisenberger et al. 1990).. In an organization, job attendance and performance of employees are positively related to their perceived support (Eisenberger et al. 1990). Further, the increased affective attachment and the expectancies of performance outcome will lead them to help the organization through some behaviors such like proposing valuable suggestion. We argued that the perceived support in brand community will strengthen the identification of the community and make the positive impact to it.. H1: Greater perceived community support is directly and positively related to greater brand community identification.. 15.

(21) 2.2.2 Community Value Congruence In addition to perceived community support, value congruence is also a antecedent of identification. Values are basic convictions of an individual which motivate action and help them to define their identities (Fombelle et al. 2012).. Two entities perceived that they have similar value is called value congruence, and they tend to perceive external stimuli in similar ways (Meglino and Ravlin 1998). It has been shown that perceived value congruence with the organization will increase job satisfaction, identification and extending the relationship with the organization ( Meglino and Ravlin 1998).. Bhattacharya and Sen (2003) argued that alike to consumers' brand congruity perceptions on self relevant dimensions, their perceptions of congruence between their own identity and that of relevant companies can be a source of self-definition. Therefore, we deemed that value congruence between a member and the brand community will strengthen his or her identification with the community.. H2: Greater perceived community value congruence is directly and positively related to greater brand community identification. 2.2.3 Community Affirmation A form of identity verification is affirmation because individuals usually seek positive self views (Drigotas et al. 1999). Affirmation is described as positive feedback that a person has reached or is reaching an identity goal. For instance, an individual may pursue the goal of being a good parent, a top sale, and a great basketball player. 16.

(22) The organization provide the affirmation of important personal identities make individual to incorporate more of their energy into the organization. Individuals will attend to act like the particular identity that have a strong connection to him/her (Arnett et al. 2003). They'll try to find organizations that provide them with such affirmation (Fombelle et al. 2012). We proposed that the members' identification with the brand community has positively related to the degree of the affirmation community provided to them.. H3: Greater perceived community affirmation is directly and positively related to greater brand community identification.. 2.3 Peer Identification Based on the research of Bhattacharya and Sen (2003), identification is influenced by three factors that are identity similarity, identity distinctiveness and identity prestige. This framework of identification is considered as a identity-related and individual-level framework because it satisfy three basic self-definitional needs (Bhattacharya & Sen, 2003). The needs of self-continuity are reflects by that individuals will find an entity which has similar identity with them. In addition, identity distinctiveness satisfy the needs of self-distinctiveness and identity prestige fulfills the needs of self-enhancement. The framework of identification which proposed by Bhattacharya and Sen (2003) is provided in figure 2.2.. It is worthy to be mentioned that the paper argues that identity of an online brand community is constructed by the members in it, not the enduring features of an entity 17.

(23) itself, such as company or brand. In other words, identity similarity, identity distinctiveness and identity prestige of an online brand community argued in this paper are come from the members inside the community. Perceived similarity with other members as well as perceived distinctiveness and perceived prestige of them in the community will attract individuals. Therefore, we apply Bhattacharya and Sen's (2003) framework to the peer-to-peer identification in the brand community; further, member-to-member identification.. Source: Bhattacharya & Sen (2003) Figure 2.1 Bhattacharya and Sen’s framework of organizational identification. 2.3.1 Peer Identity Similarity Individuals will always be attracted by someone or something that perceived to be similar factually or symbolically with their own social identity according to the similarity-attraction theory. Besides, individuals always create their self-image through products and brands they used for achieving the social identity goals (Huffman, 18.

(24) Ratneshwar and Mick, 2000). In organizational research, Pratt (1998) observed that people will identify with an organization which can fulfill their own needs for self-continuity so that they can establish their cognitively consistent social identities. Bhattacharya and Sen (2003) also claimed that when a company's identity matches consumers' identity, they will be attracted. In order to consumer psychology literature, it is verified that a product will become more desirable when the consumers feel similarity with its typical user. In this paper, we proposed that the individual's identification with other members in the brand community has positively related to the degree of the similarity they perceived between themselves and other members in it.. H4: Greater perceived member similarity is directly and positively related to greater peer identification. 2.3.2 Peer Identity Distinctiveness Tajfel and Turner (1985) argued that people will always try to make. themselves. distinctive against others in any social context they participate in. It means that individuals’ want to be unique and special. The argument however seems to be a contradiction with the sense of identity similarity. Brewer (1991) explained that people will always identify with groups which can satisfy both needs. In other words, the groups they choose to identify with would be unique from other groups and similar to their self-concept simultaneously. Besides, the groups people choose to identify with are not required to be a entirely distinctive one but particularly distinctive on the dimensions they value. In this research, we suggest that the individual's identification with other members in the brand community has positively related to the degree of the distinctiveness they perceived with the brand community against other community.. 19.

(25) H5: Greater perceived member distinctiveness is directly and positively related to greater peer identification. 2.3.3 Peer Identity Prestige A organization has a great potential to promote self-esteem through identification when it is prestigious (Mael and Ashforth, 1992). To fulfill the needs for self-enhancement, joining a prestigious organization is a common way (Ashforth & Mael, 1989; Dutton et al., 1994). Based on Bergami and Bagozzi (2000), prestige are positive perception which organization’s stakeholders’ perceive that other peoples' evaluation towards the organization. In consequence, in terms of organizational identification, Bhattacharya and Sen (2003) claimed that consumers will reflect themselves in the prestige of a brand or a company in order to raise their sense of self-worth. However, this research argues that a member in the brand community will enhance the identification with other members when they perceived the prestige of other members.. H6: Greater perceived member prestige is directly and positively related to greater peer identification. 2.3.4 Peer Identification to Brand Community Identification People’s identification with an organization is influenced not only by the relationship between individuals and the organization, but also by their identification with other individuals involved in the organization, such as other customers and employees. Prentice et al. (1994) validated the distinction between common-bond groups, which are based mainly on identification with group members, and common-identity groups, which are based mainly on direct identification to the organization. In common-bond groups, the extent to which an individual knows and 20.

(26) feels similar to other individuals of the group affect the formation of a group identification. However, in common identity groups, the identification of a group depends on one’s commitment to the identity of the group, not to the individuals within the group. Moreover, McAlexander et al. (2002) argue that important relationships can form between not only the customer and the organization, but the customer and other customers. Therefore, we suggest that the identification with other members in the brand community will also lead to higher brand community identification.. H7: Greater peer identification is directly and positively related to greater brand community identification.. 2.4 Brand Loyalty Loyalty can be viewed as a sense of attachment toward a certain set of brands and companies. Companies can obtain competitive advantages in marketing when their consumers are strongly loyal. For example, reduced marketing and transactional costs, positive word-of-mouth effect, and lowered the risk of failure (Griffin, 1996). Aaker(1991) proposed that there has six dimensions of loyalty: consumer willingness to repurchase, price premium, satisfaction rate, switching cost, preference over brand, and commitment to brand. Further, Oliver (1999) argued that consumer loyalty derived from a high level of commitment will make them repurchase the product or service.. Griffin (1996) also described loyalty as an important role in expanding e-commerce in that it can establish any kind of relationship marketing. According to the preceding research, loyalty can lead to the behaviors such like brand recommendation and retention of consumer. 21.

(27) There have been many factors influencing loyalty, such as service and product quality, brand image, price, and commitment, but few research mentioning the relationship between brand loyalty and the identification with brand community. As we mentioned before, people’s identification with an organization is influenced not only by the relationship between individuals and the organization, but also by their identification with other individuals involved in the organization. Moreover, the identification with the brand community will increase the degree of members' participation.. In Bagozzi and Dholakia's (2006) report about the purchase behaviors in a motorcycle community, members' participation in brand community(ex. browsing, discussion, on-line meeting) has a significant impact to their brand-related behavior(brand-purchase behaviors, visit the product display center). Algesheimer et al.(2005) explored that the more the individual's aspiration to maintain the membership in brand community is, the greater their loyalty toward the brand. Therefore, we suggested that both the members' identification toward the brand community and other members within can be a factor affecting the brand loyalty of members.. H8: Greater brand community identification is directly and positively related to greater brand loyalty. H9: Greater peer identification is directly and positively related to greater brand loyalty.. 22.

(28) 2.5 Moderating role of Identity Fusion Identity fusion is defined as a group identification phenomenon in which "the self-other barrier is blurred and the group comes to be regarded as functionally equivalent with the personal self (Swann et al., 2009)." This concept is about not a simple feeling particularly close to one's group, nor is it a matching closely some group prototype. For these fused people, one's group identity is one's personal identity, and vice versa.. When individuals become fused with a group, their personal and social identities become functionally equivalent and it encourages some extreme behaviors (Swann et al., 2009 ). This phenomenon is particularly common in relational group where members in it have extremely close personal relationships with one another such as family members and close friends.. In practice, a great diversity of activities and issues is more common in an online brand community. Members in this kind of brand communities just like live in a small society, doing things not only surround the brand but everything they want. It makes the connection between community and its brand more blurred and weak. Customers may join the brand community because of the fancy to the brand at first, but sustained participation in it depend on the relationship and interaction with other members. In this research, we transform the concept of identity fusion between individuals and group abovementioned into the members perception of the degree of identity fusion between the brand community and brand itself. We propose that the impact of member similarity on peer identification should be moderated by this brand-community identity fusion. An online brand community with low brand-community identity fusion is more 23.

(29) likely to strengthen the link between members by meet their various needs, and vise versa.. H10: Brand-Community identity fusion moderates the association between member similarity and peer identification: Low (versus high) brand-community identity fusion is more likely to form the peer identification in online brand communities.. 24.

(30) Chapter 3：Methodology. 3.1 Research Model This paper establishes a framework of studying the antecedents of identification in a brand community and the moderating effect of brand-community identity fusion. The whole picture of this research is shown in Figure 3.1.. Figure 3.1 Research model It is argued that peer identification in the brand community will also raise the identification to the community. Further, lead the members to higher brand loyalty. Besides, this paper examines the degree of identity fusion between the brand community and the brand which members within perceived will moderate the relationship between member similarity and peer identification. To sum up, this paper proposes several research hypotheses as follows: 25.

(31) Table 3.1 Hypotheses Hypotheses H1. H2. H3. H4. H5. H6. H7. H8. H9. Valence. Greater perceived community support is directly and positively related to greater brand community identification. Greater perceived community value congruence is directly and positively related to greater brand community identification. Greater perceived community affirmation is directly and positively related to greater brand community identification. Greater perceived member similarity is directly and positively related to greater peer identification. Greater perceived member distinctiveness is directly and positively related to greater peer identification. Greater perceived member prestige is directly and positively related to greater peer identification. Greater peer identification is directly and positively related to greater brand community identification. Greater brand community identification is directly and positively related to greater brand loyalty. Greater peer identification is directly and positively related to greater brand loyalty. Brand-Community identity fusion moderates the association. H10. between member similarity and peer identification: Low (versus high) brand-community identity fusion is more likely to form the peer identification in online brand communities.. 26. Positive. Positive. Positive. Positive. Positive. Positive. Positive. Positive. Positive.

(32) 3.2 Sampling and Data Collection Procedures Based on the survey from Foreseeing Innovative. New Digiservices (FIND) of. Institute For Information Industry (IFII) in 2011, members in brand communities would discuss the information of the product to each other.. About 80 percents of the. participant in this survey talked about the issue of education and sports, and above 60 percents of participant talked about topics about entertainment, automobiles and 3C products(see figure 3.2). In this research, we prefer to choose the communities which have not only specific brand, but also tangible products like cars.. Source: FIND of IFII (2011) Figure 3.2 Types of issues discussed through communities. As a result, we collected data from a survey of online members of nine well-known brand communities about automobiles in Taiwan. Listed below (see Table 3.2) are nine automobile brand communities.. 27.

(33) Table 3.2 Brand communities as subjects of study Types. Brand. Community Name. Automobiles. Ford. Focus-Sport Club (FSC). Nissan. SAVRIN Happy Family Club Nissan Tiida Club. Volkswagen. VW LUPO Club. Infiniti. 555 Club. Renault. Renault Fans Club. Luxgen. iLuxgen. Honda. VTEC SPIRIT. Mitsubishi. Taiwan Colt Plus Club. The reasons for choosing these brand communities: The information and knowledge exchange in these brand communities is obvious and the diversity in subjects make them flourishing. In addition, most of the registered members are discovered attend to develop the relationship with others by organizing offline meeting and activities, such as outdoor activities and tour shooting.. In order to test the research model and hypotheses, we designed a questionnaire and administered to registered members of the brand communities. This study established a web questionnaire based on the network questionnaire system provided by my3q. To avoid the repeat-respondents, the system will filter duplicate respondents by recording IP addresses. In addition, in order to enhance the motivation to respond, we have lucky draws for those who fills the valid questionnaires.. 28.

(34) 3.3 Questionnaire Design The research model of this paper consists of several constructs, including brand community. identification,. community. support,. community. value. congruence,. community affirmation, peer identification, member similarity, member distinctiveness, member prestige, brand loyalty and brand-community identity fusion.. The measures used in this research were all developed by adapting existing scales from prior studies. Necessary modifications were done in accordance with the research context of this paper. Likert scales, which consist of 7-point scale ranging from “strongly disagree”(1) to “strongly agree”(7) were used for all items.. 29.

(35) 3.4 Measures. Table 3.3 show the operational definitions and measurement items of each variable.. Table 3.3 Operational definitions and measurement items Construct/Item (Operational definitions) Brand Community Identification (Ashforth and Mael 1989) (Degree to which members see themselves as a member or belong to a brand community) 1.. When someone criticizes the (this brand community), it feels like a personal insult.. 2.. I am very interested in what others think about (this brand community).. 3.. When I talk about (this brand community), I usually say ‘we’ rather than ‘they.. 4.. (This brand community’s) successes are my successes.. 5.. When someone praises (this brand community), it feels like a personal compliment.. Community Support (Eisenberger et al. 1990) (Extent to which members' perception of the support from a brand community) 1.. (This brand community) cares about my well-being as a member.. 2.. (This brand community) cares about my opinions as a member.. 3.. (This brand community) is willing to help me in my role as a member.. 4.. (This brand community) considers my goals and values as a member.. Community Value Congruence (Maxham and Netemyer 2003) (Extent to which members' perception of the similarity of the value with a brand community) 1. This brand community has the same values as I do.. 30.

(36) Table 3.3 Operational definitions and measurement items(continue) Construct/Item (Operational definitions) Community Affirmation (Drigotas et al. 1999) (Extent to which the members' perception of the positive feedback provided by a brand community) 1.. (This brand community) sees me as a good member.. 2.. (This brand community) thinks I have the traits and dispositions of a good member.. 3.. (This brand community) treats me like I am a good member.. Peer Identification (Ashforth and Mael 1989) (Degree to which members' identification with other members in a brand community) 1.. When someone criticizes other members of (this brand community), it feels like a personal insult.. 2.. I am very interested in what others think about other members of (this brand community).. 3.. When I talk about other members of (this brand community), I usually say ‘we’ rather than ‘they.. 4.. The members’ successes are my successes.. 5.. When someone praises other members of (this brand community), it feels like a personal compliment.. Member Similarity (Shen et al.,2010 ; Turban et al, 1988; Liden et al., 1993) (Extend to which there is a match in values, interests and so on between oneself and other members of a brand community) 1.. I share similar preferences with other members of (this community).. 2.. I share similar interests with other members of (this community).. 3.. I share similar values with other members of (this community).. 4.. I joined in (this community) for the same purpose as other community members do.. 5.. The members of (this community) and I see things in a similar way.. 6.. The members of (this community) and I deal with issues in much the same way.. 31.

(37) Table 3.3 Operational definitions and measurement items(continue) Construct/Item (Operational definitions) Member Distinctiveness (Jones and Volpe, 2010) (Extend to which individuals perceive other members in a brand community to be unique and different from the out-group) 1.. When I think about (this community), the way it runs the community is unique and different from other communities.. 2.. The other members of (this community) are unique in my point of view.. 3.. There are some experts who only participate in (this community).. 4.. The purposes of (this community) outshine the other communities.. 5.. The members of (this community) have special and unique characteristics.. Member Prestige (Ashforth and Mael, 1992) (Extend to which individuals perceive other members in a brand community to be prestigious) 1.. Experts generally think highly of (this community).. 2.. It is considered prestigious in (the community) of fans to be a member of this community.. 3.. This community is considered one of the best (in this kind of communities).. 4.. People from other communities respect the members of (this community).. 5.. Members of (this community) would be proud to introduce others the other members of this community.. 6.. The members of (this community) have a good reputation.. 7.. Showing associations with (this community) could make oneself respected.. 8.. Members of (this community) are widely welcome by any kinds of contest and activities.. 32.

(38) Table 3.3 Operational definitions and measurement items(continue) Construct/Item (Operational definitions) Brand Loyalty (Chaudhuri and Holbrook, 2001) (Extend to which members in a brand community willing to purchase or use the products of the brand) 1.. I will buy products of this brand next time.. 2.. I prefer to keep buying products of this brand.. 3.. I am loyalty to this brand.. 4.. I would pay higher price to this brand rather than other brand.. Identity Fusion(Swann, Gómez, Huici, Morales and Hixon, 2010) (The degree of the fusion between the community and the brand member perceived) I think the degree of brand-community fusion is:. Community Brand. A. Community Brand. B. Community Brand. C. Community Brand. D. 33. Community Brand. E.

(39) Chapter 4: Results 4.1 Descriptive Statistics A total of 420 respondents representing different automobile brand communities collected in the survey. However, 89 responses were found to be invalid as they were either miss-responses or not fit in our objective communities of this study, and thus eliminated to ensure the quality of the survey results. 51 responses were afterwards filtered out due to answering with stereotyped pattern and failed to meet the logical consistencies. Only 280 responses were left valid for follow-up analysis at last. 4.1.1 Sample Characteristics Questionnaire was mainly administered online through posting URL address of online web questionnaire in PTT, Facebook and the automobile brand communities. Of the entire sample, only 3.9% were female, the other 96.1% were male. This result shows that members of automobile brand communities are predominantly men.. Respondents ranged in age from 20 to above 50, with most of the respondents come from the age range of 30-35 (34.3%). While 83.3% were working people, the other 2.1% were students and the remained 14.6% did not specify.. With more than half of the respondents was either undergraduate or postgraduate degree holders (67.5%), the respondents of this survey were believed to be relatively well-educated and knowledgeable. Table 4.1 summarizes the respondent characteristics.. 34.

(40) Table 4.1 Descriptive statistics of respondents’ characteristics (N=280) Characteristics Gender. Age. Occupation. Education. Annual Income (in thousands). Items. Frequency. Percentage. Male Female ＜20~25 ＜25~30 ＜30~35 ＜35~40 ＜40~45 ＜45~50. 11 269. 3.9% 96.1%. 22 76 96 45 26 8. 7.9% 27.1% 34.3% 16.1% 9.3% 2.9%. ＜50 and above. 7. 2.5%. Housework Military, civil and teaching Service industry Banking and insurance industry IT industry Manufacturing and transportation Business Freelance SOHO. 1 31 36 9 51 79 13 12 1. 0.4% 11.4% 12.9% 3.2% 18.2% 28.2% 4.6% 4.3% 0.4%. Student Others. 6 41. 2.1% 14.6%. High school Technical institute Undergraduate Postgraduate and above. 35 56 109 80. 12.5% 20.0% 38.9% 28.6%. 300 or below ＜300~400 ＜400~500 ＜500~600. 22 34 46 41. 7.9% 12.1% 16.4% 14.6%. ＜600~700 ＜700~800 ＜800~900 ＜900~1000 ＜1000 and above. 33 34 16 17 37. 11.8% 12.1% 5.7% 6.1% 13.2%. 35.

(41) Table 4.1 Descriptive statistics of respondents’ characteristics (N=280)(continue) Yearly Expenses on Automobile Equipments (in thousands). 5 or below ＜5~10 ＜10~15 ＜15~20 ＜20~25 ＜25 and above. 11 42 49 23 29 126. 3.9% 15.0% 17.5% 8.2% 10.4% 45.0%. 4.1.2 User Behavior Analysis Figure 4.1 and figure 4.2 shows the types of car possessed and the car most-used by survey respondents respectively. Because questionnaire was mostly distributed in FSC, FORD ranked highest in both results. Besides, figure 4.3 indicates respondents' usage of their car, most of them have used their own car for 5 years. In addition, these 3 questions served not only as contingency questions but also sleeper questions, which used to search out for abnormality.. Referring to the statistical results from figure 4.4 and figure 4.5, Focus-Sport Club(FSC) is the most popular automobile brand community perhaps due to the abovementioned reason. In addition, these 2 questions were also sleeper questions.. 36.

(42) Survey respondents' possession of car (Multiple-choice) N=280 200 180 160 140 120 100 80 60 40 20 0. 176. 72 53 14. 15. 11. 20. 25. Figure 4.1 Statistical result of survey respondents’ possession of automobile. Survey respondents' most-used car N=280 160 140. 151. 120 100 80 60 40 20 0. 12. 32. 4. 11. 20. 10. 11. 10. Figure 4.2 Statistical result of survey respondents’ most-used car. 37. 19.

(43) Survey respondents' length of car usage N=280 90 80 77. 70 60 59. 50 40 30 20. 29. 25. 10 0. 17. 16. 8. 18. 12. 15. 4. Figure 4.3 Statistical result of survey respondents’ length of car usage. Survey respondents' membership of automobile brand communities(Multiple-choice ) N=280 200 180 160 140 120 100 80 60 40 20 0. 176. 12. 2. 17. 6. 21. 4. 21. 13. Figure 4.4 Statistical result of survey respondents’ membership of automobile brand communities. 38. 46.

(44) Survey respondents' most-visited automobile brand communities N=280 180 160 140 120 100 80 60 40 20 0. 160. 11. 14. 21. 4. 2. 20. 12. 36. Figure 4.5 Statistical result of survey respondents’ most-visited automobile brand communities. There were several questions designed to reveal the user behavior of the brand communities’ members. These questions included car usage, membership details, extent of community participation and so on.. Table 4.2 summarizes the user behavior of respondents. Most of them has been registered members ranged from 6 months to 2.5 years as well as 87.5% of the respondents visited automobile brand communities more than 11 times a month. Besides, about 50% of the respondents browsed the communities for more than 15 minutes but less than 45 minutes. This shows that most of the respondents are remarkably fit for analysis.. As most of the respondents (66.5%) indicated that they either often or always interacted with other members, showing the evidence of high level of interaction among members of automobile brand communities. However, with 80.8% of the respondents 39.

(45) indicated that they never/seldom/sometimes published and/or replied photos and/or articles, they were viewed to be less contributive as they might be behaving as a lurker most of the time. In last, most of the respondents (91.1%) indicated that they never/seldom/sometimes participate in activities held by the community, this suggests that the participation in automobile brand communities is frequent online more than offline.. Table 4.2 Analysis of user behavior (N=280) Category. Length of membership. Frequency of logging-on (month). Browsing duration (in minutes). Awareness of others on membership. Items. Frequency. Percentage. 6 months or below <6 months~1 year <1 year~1.5 years <1.5 years~ 2 years <2 years~2.5 years <2.5 years~3 years <3 years~3.5 years <3.5 years~4years. 71 64 0 55 23 0 22 12. 25.4% 22.9% 0% 19.6% 8.2% 0% 7.9% 4.3%. <4 years~4.5 years <4.5 years~5 years <5 years and above. 0 8 25. 0% 2.9% 8.9%. 1~2 3~4 5~6 7~8 9~10 11. 4 6 9 8 8 245. 1.4% 2.1% 3.2% 2.9% 2.9% 87.5%. 15 or below. 30. 10.7%. <15~30 <30~45 <45~60 <60~75 <75~90 <90 and above. 98 58 24 21 11 38. 35.0% 20.7% 8.6% 7.5% 3.9% 13.6%. No Yes. 166 114. 59.3% 40.7%. 40.

(46) Table 4.2 Analysis of user behavior (N=280)(continue). Interaction among members. Never Seldom Sometimes Often Always. 7 29 58 136 50. 2.5% 10.4% 20.7% 48.6% 17.9%. Contributions (publishing and/or replying photos and/or articles). Never Seldom Sometimes Often Always. 24 54 148 44 10. 8.6% 19.3% 52.9% 15.7% 3.6%. Participation in activities held by communities. Never Seldom Sometimes Often Always. 111 71 73 19 6. 39.6% 25.4% 26.1% 6.8% 2.1%. 41.

(47) 4.2 Factor Analysis Because most of the measurements used in this research were previously conducted in offline and working context, we modified some items to fit the research context. We perform a factor analysis in order to check on the validity and reliability as well as unidimensionality of all variables in this research because each variable consisted of multiple items.. First of all, we use a confirmatory factor analysis (CFA) to examine our measurements and deleted several items because of unacceptably low loadings or high cross-loadings. The results of the confirmatory factor analysis (CFA) indicated that the measurement model of our research was well fit with the data. Table 4.3 shows that the convergent validity of all constructs was satisfied. The CFA factor loadings for all items were respectable and the Cronbach’s alpha values for all constructs were near to or greater than the conventional value of 0.7, showing high internal consistency.. The summary of the latent construct’s composite reliability, average variance extracted (AVE), and between-construct correlations shows in table 4.4. Seeing table 4.4, the estimates of composite reliability were all greater than .85 and showed a high level of internal consistency (Bagozzi and Yi 1988; Fornell and Larcker 1981). In other hand, all the square roots of the AVEs on the diagonal were greater than the off-diagonal elements, shows the evidence of discriminant validity.. In addition, the AVE values for all constructs were significantly higher than the stipulated criteria (50%), suggesting well validity of each latent variable. In sum, the evidence for good convergent validity are proved by all the tests. Also, there was 42.

(48) evidence of an adequate level of discriminant validity too. Table 4.5 providing the evidence of discriminant validity, and all the construct intercorrelations were significantly less than 1 (p<.05).. In consequence, all data were good to analysis in validity and reliability. Good result of convergent validity and discriminant validity allow us move to next step: estimate the structural model.. 43.

(49) Table 4.3 Factor analysis results Construct Community Support. ID. Measurement item. Mean. SD. CFA factor. Cronbach‘s. loading. α 0.87. 0.78. 0.94. 0.90. 0.91. 0.70. 0.92. 0.77. V19. 1.. (This brand community) cares about my well-being as a member.. 5.07. 1.238. 0.89. V20. 2.. (This brand community) cares about my opinions as a member.. 5.13. 2.173. 0.92. V22. 3.. (This brand community) considers my goals and values as a. 4.98. 1.181. 0.84. AVE. member. Community. V24. 1.. This brand community) sees me as a good member.. 5.05. 1.107. 0.95. Affirmation. V25. 2.. This brand community) thinks I have the traits and dispositions of. 4.96. 1.080. 0.95. a good member. Member Similarity. V26. 3.. This brand community) treats me like I am a good member.. 5.10. 1.064. 0.94. V34. 1.. I share similar preferences with other members of (this. 5.28. 1.104. 0.82. community). V35. 2.. I share similar interests with other members of (this community).. 5.17. 1.138. 0.86. V36. 3.. I share similar values with other members of (this community).. 4.88. 1.247. 0.86. V37. 4.. I joined in (this community) for the same purpose as other. 5.19. 1.133. 0.81. 4.74. 1.146. 0.86. 4.71. 1.197. 0.81. 4.87. 1.223. 0.84. community members do. V38. 5.. The members of (this community) and I see things in a similar way.. V39. 6.. The members of (this community) and I deal with issues in much the same way.. Member Distinctiveness. V40. 1.. When I think about (this community), the way it runs the community is unique and different from other communities. 44.

(50) V41. 2.. The other members of (this community) are unique in my point of. 4.73. 1.203. 0.91. view. V42. 3.. There are some experts who only participate in (this community.. 5.22. 1.268. 0.82. V43. 4.. The purposes of (this community) outshine the other. 4.94. 1.150. 0.91. 4.90. 1.178. 0.89. communities. V44. 5.. The members of (this community) have special and unique characteristics.. Member Prestige. V45. 1.. Experts generally think highly of (this community).. 5.01. 1.136. 0.68. V46. 2.. It is considered prestigious in (the community) of fans to be a. 5.02. 1.257. 0.80. 5.24. 1.138. 0.68. 4.98. 1.151. 0.68. 5.23. 1.121. 0.70. 0.94. 0.52. NA. 0.83. member of this community. V47. 3.. This community is considered one of the best (in this kind of communities).. V48. 4.. People from other communities respect the members of (this community).. V49. 5.. Members of (this community) would be proud to introduce others the other members of this community.. V50. 6.. The members of (this community) have a good reputation.. 5.06. 1.093. 0.76. V51. 7.. Showing associations with (this community) could make oneself. 4.87. 1.258. 0.72. 4.98. 1.173. 0.73. 5.08. 1.433. 0.91. 4.54. 1.490. 0.91. respected. V52. 8.. Members of (this community) are widely welcome by any kinds of contest and activities.. Brand Community. V16. 1.. Identification. When I talk about (this brand community), I usually say "we" rather than "they".. V17. 2.. (This brand community’s) successes are my successes. 45.

(51) Peer Identification. V29. 1.. When someone criticizes other members of (this brand. 4.38. 1.429. 0.72. 4.59. 1.401. 0.88. 5.04. 1.284. 0.87. 0.86. 0.68. NA. 0.89. community), it feels like a personal insult. V32. 2.. The members’ successes are my successes.. V33. 3.. When someone praises. other members. of. (this. brand. community), it feels like a personal compliment. Brand Loyalty. V54. 1.. I prefer to keep buying products of (this brand).. 5.39. 1.268. 0.94. V55. 2.. I am loyalty to (this brand).. 4.95. 1.355. 0.94. Table 4.4 Correlations among constructs, convergent validity and discriminant validity Composite reliability. Average variance extracted. 1. Brand community identification. 2(5). 0.91. 0.83. 2. Peer identification. 3(5). 0.86. 0.68. Correlation between 1. Latent 0.91 Constructs 0.79. 3. Brand loyalty. 2(4). 0.94. 0.89. 0.40. Construct. Number of items. 46. 2.. 0.82 0.42. 3.. 0.94.

(52) 4.3 Structural Equation Model In structural equation model, we estimate the hypothesized conceptual model shown in the figure 3.1. 4.3.1 Test of Goodness-of-fit The chi-square value was not significant (χ2(279) = 63.747, p>0.05) and all the statistics were good, indicating a good fit to the data. Table 4.5 reports the respectively absolute fit measures, comparative fit measures and parsimonious fir measures, and the overall goodness-of-fit indices are satisfied.. Table 4.5 Goodness-of-fit statistics for conceptual model Index. Value. Absolute fit χ2/d.f. measures GFI. Comparative fit measures. Parsimonious fit measures. Stipulated criteria. 1.45 0.966. <3 >0.9 (Joreskog and Sorbom, 1996) >0.9 (Joreskog and Sorbom, 1996) The smaller, the better (Hair, 1998) <0.05 (Browne and Cudeck, 1993). AGFI. 0.930. RMR. 0.0289. RMSEA. 0.0401. NFI. 0.989. NNFI. 0.993. >0.9 (Bentler and Bonnett, 1980) >0.9. CFI IFI RFI. 0.996 0.996 0.980. (Tucker and Lewis, 1973) >0.9(Bentler, 1990) >0.9(Bentler, 1990) >0.9(Hu and Bentler, 1999). PNFI. 0.558. PGFI. 0.467. >0.5 (Joreskog and Sorbom, 1996) >0.5 (Joreskog and Sorbom, 1996) 47. Result Good Good Good Good Good Good Good Good Good Good Good Acceptable.

(53) 4.3.2 Parameter estimation There were 6 exogenous variables used in this research: community support (ξ1), community value congruence (ξ2), community affirmation (ξ3), member similarity (ξ4), member distinctiveness (ξ5) and member prestige (ξ6). On the other hand, 3 endogenous variables were used in this research: brand community identification (η1), peer identification (η2), brand loyalty (η3). Figure 4.6 illustrates the path diagram.. The standardized estimates on latent endogenous variables and observed variables are shown in table 4.6, including the results of SMC and t-value for each construct and item. The t-value of all items were statistically significant (t>3.29, p<0.001), indicating a good quality in terms of measurement characteristics, all SMC values were also satisfactory simultaneously.. Figure 4.6 Path diagram 48.

(54) Table 4.6 Standardized estimates on latent endogenous variables and observed variables λy. Observed variables. Standardized coefficient. t-value. Brand Community Identification (η1). Y1. 0.73. ─. 0.49. Y2. 0.89. 12.33***. 0.73. Y3. 0.71. ─. 0.46. Y4. 0.48. 4.88***. 0.73. Y5. 0.91. 12.26***. 0.76. Y6. 0.84. ─. 0.65. Y7. 0.92. 8.82***. 0.79. Peer Identification (η2 ) Brand Loyalty (η3) p<0.05;. ＊. SMC. 0.73. 0.68. 0.16. p<0.01; ***p<0.001. ＊＊. Table 4.7 and 4.8 summarize the standardized effects for the conceptual model. Relationships between endogenous variables are illustrated by β value as well as the relationships between exogenous variables and endogenous variables are illustrated by γ value. Referring to table 4.7 and 4.8, all paths are statistically significant in addition to H2 and H8. Thus, community value congruence (ξ2) is found to have indirect effect (γ12=0.02) on brand community identification (η1). Even more surprising is that brand community identification (η1) is also found to have indirect effect (β31=0.07) on brand loyalty (η3). Moreover, there is an unexpected discovery that showing that brand community support (ξ1) has direct effect (γ22=0.24) on peer identification (η2).. 49.

(55) Table 4.7 Decomposition of standardized effects for conceptual model (γ value) γ. Path. Valence. Path coefficient. t-value. Result. γ11. ξ1→η1. 0.28. 3.75***. H1 supported. γ12. ξ2→η1. 0.02. 0.38. H2 unsupported. γ13. ξ3→η1. 0.16. 2.31*. H3 supported. γ24. ξ4→η2. + + + + + + +. 0.29. 4.59***. H4 supported. 0.17. 2.60**. H5 supported. 0.27. 3.77***. H6 supported. 0.24. 4.15***. γ25. ξ5→η2. γ26. ξ6→η2. γ22. ξ2→η2. p<0.05;. ＊. p<0.01; ***p<0.001. ＊＊. Table 4.8 Decomposition of standardized effects for conceptual model (β value) β. Path. Valence. Path coefficient. t-value. Result. β12. η2→η1. +. 0.50. 5.56***. H7 supported. β31. η1→η3. +. 0.07. 0.53. H8 unsupported. β32. η2→η3. +. 0.50. 3.15**. H9 supported. p<0.05;. ＊. p<0.01; ***p<0.001. ＊＊. We have proposed that brand-community identity fusion would moderate the relationship between member similarity and peer identification. To test this moderating effect, we divided the participants into three groups, according to their scores on the brand-community identity fusion. We took the low degree of identity fusion (the first group) with the high degree and conduct a multi-sample test─94 participants who scored low on brand-community identity fusion (mean=2.43), and 94 participants who scored high (mean=4.85).. This research conducted a multiple group analysis to examine the moderating 50.

(56) effects. In addition, we tested the structural model with free parameter estimates and a model with an equality constraint imposed on the path between member similarity and peer identification simultaneously. A higher chi-square indicated a poorer fit, would prove a significant difference between the models for the high versus low brand-community identity fusion groups. Table 4.9 shows the moderating effect of brand-community identity fusion is significant (Δχ2=5.94, Δdf=1). The estimated coefficient of member similarity on peer identification decreased from .39 (low brand-community identity fusion, t=3.91) to .17 (high brand-community identity fusion, t=1.07). That is, the direct effect of member similarity on peer identification appeared more evident for consumers who scored low in brand-community identity fusion.. Table 4.9 Moderating effect of brand-community identity fusion Hypothesized relationship. Group. Standardized coefficient. Δχ2. Δdf. Member similarity→Peer. High identity. 0.17. 5.94. 1. identification. fusion Low identity fusion. 0.39***. 51.

(57) 4.4 Hypothesis Testing All SEM result were obtained by using maximum likelihood method. Considering the results from table 4.7, 4.8 and 4.9, 8 proposed hypotheses receive considerable amount of supports and 2 proposed hypotheses don't. Table 4.10 summarizes the results of hypothesis testing.. In terms of the antecedents of brand community identification, the results prove that community support (H1; γ=0.28, p<0.001) and community affirmation (H3; γ=0.16, p<0.05) have positively direct effects on brand community identification. However, community value congruence (H2; γ=0.02, p<0.001) isn't significant may due to the only one question as the measurement.. By comparing the standardized coefficients, member similarity has the strongest effect (H4; γ=0.29, p<0.001) over member prestige (H6; γ=0.27, p<0.001) and member distinctiveness (H5; γ=0.17, p<0.01) on peer identification. In addition, member distinctiveness is the weakest antecedent, indicating that distinctiveness among members is not the vital account of peer identification. However, we found an unexpected relationship (γ=0.24, p<0.001) between community support and peer identification. The finding indicated that community support can also directly positively influence peer identification, and it's never previously identified by any researches.. Furthermore, the results suggest that peer identification has a positively strong effects on brand community identification (H7; β=0.50, p<0.001) and brand loyalty (H9; β=0.50, p<0.01). On the other hand, contrast with our expectation, brand community identification has no direct effect on brand loyalty (H8; β=0.07). 52.

(58) In last, with regard to the moderating role of brand-community identity fusion, we discovered that low brand-community identity fusion would strengthen the positive effect between member similarity and peer identification. Path diagram with parameter estimates is illustrated in the figure 4.7.. 53.