A moderated mediation model of consumers role behaviors in brand communities

Description
The aim of this study was to investigate the manner in which individuals identify with a voluntary organization
by developing a moderated mediation model of members' role behaviors in brand community
(BC) settings. The influence of BC characteristics, including the social, emotional, and cognitive dimensions
of these characteristics, on members' identification with a BC was investigated. The empirical
results indicated that social and emotional characteristics exert a more significant influence on members'
community identification than cognitive characteristics do. Moreover, the length of the relationship
exerts a u-shaped curvilinear effect on the positive relationship between identification and extra-role
behaviors, implying that senior members experience three phases, namely honeymoon, inertia, and
rejuvenation periods, when participating in BCs. We conclude by presenting a discussion of the research
and the managerial implications of this study's findings.

A moderated mediation model of consumers' role behaviors in brand
communities
Hsiu-Yu Hung
*
, Ting-Ling Lin
Department of Business Administration, National Taipei University, Taiwan
a r t i c l e i n f o
Article history:
Received 2 March 2014
Accepted 2 December 2014
Available online 23 July 2015
Keywords:
Brand community (BC) identi?cation
In-role behavior
Extra-role behavior
Brand community characteristics
Relationship length
a b s t r a c t
The aim of this study was to investigate the manner in which individuals identify with a voluntary or-
ganization by developing a moderated mediation model of members' role behaviors in brand community
(BC) settings. The in?uence of BC characteristics, including the social, emotional, and cognitive di-
mensions of these characteristics, on members' identi?cation with a BC was investigated. The empirical
results indicated that social and emotional characteristics exert a more signi?cant in?uence on members'
community identi?cation than cognitive characteristics do. Moreover, the length of the relationship
exerts a u-shaped curvilinear effect on the positive relationship between identi?cation and extra-role
behaviors, implying that senior members experience three phases, namely honeymoon, inertia, and
rejuvenation periods, when participating in BCs. We conclude by presenting a discussion of the research
and the managerial implications of this study's ?ndings.
© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier
Taiwan LLC. All rights reserved.
1. Introduction
Effective brands must be identi?able to customers. Establishing
a loyal customer base is crucial to the sustainability of a brand. In
practice, establishing an ongoing membership in a brand commu-
nity (BC) is a considerably effective approach for gaining customer
loyalty (Algesheimer, Dholakia, & Herrmann, 2005). A BC is a
specialized organization centered on a brand and characterized by a
shared consciousness, rituals and traditions, and a sense of moral
responsibility (Muniz & O'guinn, 2001). A BC is based on a struc-
tured set of social relationships among brand admirers whose
shared enthusiasm for the target brand manifests a well-developed
social identity that expresses shared sentiments and encourages
members' participation in achieving collective goals (Bagozzi &
Dholakia, 2006). Therefore, BCs are highly valuable for marketing
practitioners in brandeconsumer relationship management.
Numerous scholars have invested considerable efforts in un-
derstanding members' participation according to both the factors
that precede participation and those that mediate it (e.g.,
Algesheimer et al., 2005; Bagozzi & Dholakia, 2006; Dholakia,
Bagozzi, & Pearo, 2004; Pai & Tsai, 2011; Tsai & Pai, 2013). One of
the most consistent ?ndings is that social identi?cation with a
community plays a crucial role. Identi?cation refers to the degree to
which members perceive similarities between their de?ning at-
tributes and those of the focal organization (Bhattacharya & Sen,
2003; Scott & Lane, 2000). A member's identi?cation with a com-
munity acts as an essential mediating mechanismthat shapes his or
her proactive participation (Ahearne, Bhattacharya, & Gruen, 2005;
Algesheimer et al., 2005; Bagozzi &Dholakia, 2006; Bhattacharya &
Sen, 2003; Tsai & Pai, 2013). Companies and BC practitioners thus
favor these behaviors. Despite their importance, the antecedents of
member participation and conditions under which identi?cation
in?uences members' participation and decision making have not
been addressed thoroughly in the literature. The objective of this
study was to remedy this oversight and address the following
related research gaps.
The ?rst gap concerns the antecedents of BC identi?cation from
the perspective of the characteristics of a BC. Although Tsai and Pai
(2013) introduced the relevance of community characteristics as
antecedents of proactive participation in virtual community con-
texts, the factors that precede a member's identi?cation with a BC
are not clearly understood. However, their investigation of partic-
ipation antecedents included only a limited number of constructs,
thus necessitating further research into other factors that in?uence
* Corresponding author. Department of Business Administration, National Taipei
University, New Taipei City, 23741, Taiwan.
E-mail address: [email protected] (H.-Y. Hung).
Peer review under responsibility of College of Management, National Cheng
Kung University.
HOSTED BY
Contents lists available at ScienceDirect
Asia Paci?c Management Review
j ournal homepage: www. el sevi er. com/ l ocat e/ apmrvhttp://dx.doi.org/10.1016/j.apmrv.2014.12.014
1029-3132/© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier Taiwan LLC. All rights reserved.
Asia Paci?c Management Review 20 (2015) 191e200
members' participation to understand clearly the factors that in-
?uence members to identify with a BC.
Another research gap relates to the consequences of BC identi-
?cation from the perspective of members' in-role and extra-role
behaviors. The existing literature has not yet addressed this topic
adequately. Although researchers have examined proactive (extra-
role) (Tsai & Pai, 2013) and reactive (in-role) participation behav-
iors independently (Cheung & Lee, 2009; Koh, Kim, Butler, & Bock,
2007; Wiertz & de Ruyter, 2007), the literature lacks an empirical
exploration of situations in which both participation behaviors are
present. The current study addressed this research gap by incor-
porating in-role and extra-role behaviors into one conceptual
model.
The third gap relates to the u-shaped curvilinear effect of rela-
tionship length on identi?cation behavior. Mael and Ashforth
(1992) argued that identi?cation increases rapidly in the early
stages of a relationship (the honeymoon period), followed by a
period of inertia, indicating that tenure with an organization
weakens the positive relationship between identi?cation and role
behaviors. Furthermore, Tsai and Bagozzi (2014) argued that as
members participate in social groups over time, the possibility that
their positive attitudes transform into positive behaviors increases,
implying that the length of the relationship strengthens the posi-
tive relationship between identi?cation and role behaviors. To the
best of our knowledge, BC studies lack consistent evidence detailing
the role of members' tenure in members' participation processes.
To address this gap, we adopted the perspectives of Mael and
Ashforth (1992) as well as Tsai and Bagozzi (2014) and propose
that individuals' participation experience comprised three phases:
honeymoon, inertia, and rejuvenation periods. This proposition
represents a considerable ?rst attempt to examine the u-shaped
nonlinear effect of relationship length on members' identi?cation
and role behaviors in BC settings.
2. Theoretical background and hypothesis development
The proposed research framework (Fig. 1) elaborates on the
nature of BC identi?cation. We propose that social, emotional, and
cognitive characteristics of a BC affect members' identi?cation with
a community. In accordance with previous research, this frame-
work positions identi?cation as having positive consequences,
causing members to become active supporters of the BC with which
they identify. The proposed model also recognizes that, although
identi?cation may guide in-role and extra-role behaviors, rela-
tionship length can moderate this relationship.
2.1. Brand community identi?cation
Identi?cation is de?ned as the degree to which members of an
organization perceive themselves and an organization as having
the same de?ning attributes (Bhattacharya & Sen, 2003). The
central assertion in this ?eld is that some of the strongest mem-
berecommunity relationships are those that are anchored on the
consumers' identi?cationwith a community that helps themsatisfy
one or more self-de?ned needs (Ahearne et al., 2005; Algesheimer
et al., 2005). Such identi?cation is active, selective, and volitional on
the part of the members and causes them to engage in favorable
community-related behaviors (Bagozzi & Dholakia, 2006). This
assertion is supported by research that identi?es organizations as
forming key components of people's social identities. An individual
engages in self-categorization by comparing his or her personal
de?ning characteristics with those that de?ne a BC (Bagozzi &
Dholakia, 2006).
2.2. Antecedents of brand community identi?cation
An individual's identi?cation with an organization depends on
his or her perceptions of the identity of the organization as repre-
sented in its core or de?ning characteristics (Dutton, Dukerich, &
Harquail, 1994). The identity of an organization, like the identity
of an individual, represents constellations of characteristics or traits
that are central to the organization (Scott &Lane, 2000). On basis of
the ?ndings of Dholakia et al. (2004), McAlexander, Schouten, and
Koenig (2002), and Tsai and Pai (2013), this study proposed three
dimensions of BC characteristics that shape members' identi?ca-
tion with a BC: social, cognitive, and emotional dimensions.
Fig. 1. Conceptual framework.
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 192
2.2.1. Social dimension: social bene?ts
The social bene?ts of a BC include the social support, friendship,
and intimacy that are derived froma member being in contact with
other people (Dholakia et al., 2004). Members frequently join BC
activities to dispel their loneliness by meeting other like-minded
individuals who can offer companionship and support (McKenna
& Bargh, 1999). Moreover, by gaining the acceptance and
approval of others, members can enhance their social status within
a BC through their contributions to the BC (Baumeister, 1998).
Maintaining connections with others and achieving social
enhancement are the social bene?ts derived from participating in
BCs. Thus, H1: Social bene?ts positively in?uence members' iden-
ti?cation with a BC.
2.2.2. Social dimension: brand community prestige
The prestige of BCs, which is a social characteristic, is related to
the desire of people to perceive themselves in a positive light
(Bhattacharya & Sen, 2003). Kunda (1999) suggested that self-
esteem is enhanced by people's need for self-enhancement or the
maintenance and af?rmation of positive self-views. In addition,
studies have indicated that consumers use their identi?cation with
social groups that have prestigious identities as a crucial mecha-
nism for satisfying both their self-enhancement and self-esteem
needs (Dutton et al., 1994; Mael & Ashforth, 1992). When mem-
bers identify with the prestigious identity of a BC, it enables themto
view themselves as sharing in the desirable status of the organi-
zation (Bhattacharya & Sen, 2003). Members develop pride from
participating in a BC and perceive this pride as socially valuable
(Dutton et al., 1994). Thus, H2: BC prestige positively in?uences
members' identi?cation with a BC.
2.2.3. Cognitive dimension: informativeness
Informativeness is a cognitive feature that expresses the value
derived from accomplishing a predetermined instrumental pur-
pose, such as providing or receiving useful information through BC
participation (Dholakia et al., 2004). When a BC serves as an
essential source of information to help its members make decisions
or accomplish tasks, this support has crucial implications for
developing the memberecommunity relationship. Dholakia et al.
(2004) suggested that when members consider the information
they gain useful, they develop a deeper attachment to the social
group. Wasko and Faraj (2000) discovered that when members
receive valuable information from their communities, they recip-
rocate by demonstrating their identi?cation with the community.
Thus, H3: Informativeness positively in?uences members' identi-
?cation with a BC.
2.2.4. Emotional dimension: enjoyment and entertainment
Enjoyment is an emotional characteristic and is similar to the
notion (propounded by environmental psychologists) that a plea-
surable experience produces an emotional response (Cheung &Lee,
2009; Novak, Hoffman, & Yung, 2000). Brand community partici-
pants often explore various ?ctional identities as well as encounter
and then solve challenges as a means of entertainment (McKenna &
Bargh, 1999). Nambisan and Baron (2009) asserted that the hedonic
bene?ts of prior community participation facilitate attachment and
belonging to the community (i.e., BC identi?cation). Moreover,
Ahearne et al. (2005) argued that identi?cationwith a BC is possible
only when individual members perceive the community to be
attractive; therefore, the degree of identi?cation varies with in-
dividuals' perception of their levels of enjoyment. Likewise,
Rothaermel and Sugiyama (2001) reported that the integration of
members within a community is a function of the degree of plea-
sure that they experience during their engagement processes. Thus,
H4: Enjoyment and entertainment positively in?uence members'
identi?cation with a BC.
2.2.5. Emotional dimension: escapism
Escapism can be conceptualized as the manner in which in-
dividuals can refocus their attention by creating fantasies or “con-
structed unrealities” that liberate them from individual realities
induced by mass-market societies and other life stresses (Henry &
Caldwell, 2006). Henry and Caldwell (2006) discussed escapism as
a type of self-empowerment that consumers use to cope with
feelings of powerlessness triggered by engaging in the marketplace,
the nightmare of repetition, and work stress experienced in
everyday life. Individuals can temporarily escape from social stress
and indulge in a sense of freedomand pleasure by participating in a
BC (Rojek, 1993). During this process, feelings of playfulness and
?owmay be evoked, and these positive experiences lead to positive
associations and identi?cation with the BC (Hausman & Siekpe,
2009). Accordingly, H5: Escapism positively in?uences members'
identi?cation with a BC.
2.3. Consequences of brand community identi?cation
Once individuals identify with an organization, they can support
the organization in various manners. For example, individuals are
motivated to expend efforts voluntarily for an organization, commit
themselves to achieving the goals of the organization, and posi-
tively cooperate with members of the organization (Bhattacharya &
Sen, 2003). Studies have reported that members' identi?cationwith
an organization results in lower turnover (Bergami & Bagozzi,
2000; Dutton et al., 1994) and even in increased derogation of
competitors. Therefore, understanding how and why members
develop a sense of identi?cation with an organization clearly has
several crucial consequences and implications for studying BCs.
2.3.1. In-role behavior
Ahearne et al. (2005) reported that stronger levels of identi?-
cation are associated with a substantial display of in-role behaviors
that are supportive of the organization. In-role behaviors are
considered formal requirements by an organization and are directly
linked to extrinsic rewards and sanctions (Morrison, 1994). From a
social identity standpoint, once a customer identi?es with a BC,
purchasing the products of that brand becomes an act of self-
expression (Ahearne et al., 2005). In other words, members are
likely to expend speci?c efforts to support the existing products of
the brand and to try its new products (Bhattacharya & Sen, 2003).
Thus, H6: BC identi?cation positively in?uences members' in-role
behaviors in BCs, including those regarding brand choice and
product use.
2.3.2. Extra-role behaviors
Individuals who identify strongly with an organization are likely
to focus on tasks that bene?t the entire organization rather than
pursue individual self-interest (Ahearne et al., 2005). Extra-role
behaviors are behaviors that are voluntarily exerted and oriented
toward helping others (Dutton et al., 1994; Wuyts, 2007). In-
dividuals engage in extra-role behaviors such as organizational
citizenship, prosocial, spontaneous, contextual (Borman &
Motowidlo, 1993), proactive (Tsai & Pai, 2013), and contribution
behaviors (Tsai & Bagozzi, 2014) as well as transmit messages
through word of mouth; for example, they suggest product im-
provements, recruit other customers, and engage in proactive
communication (Gruen, Summers, & Acito, 2000). Members' sup-
port for a social group can be expressed through avenues other than
merely consumption, such as the promotion of an organization to
signi?cant others (Scott & Lane, 2000). Another effective path to
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 193
the long-term success of a BC, from the perspectives of members,
lies in recruiting new members to the community (Bhattacharya &
Sen, 2003). Thus, member recruitment might be a key manifesta-
tion of members' voluntary efforts to contribute to the long-term
welfare of an organization. Thus, H7: BC identi?cation positively
in?uences the extra-role behaviors of members in BCs.
2.4. U-shaped moderating effect of relationship length
Relationship length is another factor that may exert a moder-
ating in?uence on the relationship between identi?cation and
desirable role behaviors. In the ?eld of BCs, there are two per-
spectives regarding the moderating effect of the length of a rela-
tionship on community identi?cation and the behaviors of
members. Bhattacharya, Rao, and Glynn (1995) proposed that
membership length increases identi?cation at ?rst; however, the
rate of this increase diminishes over time. A newcomer to the social
context of a BC is likely to experience an initial honeymoon period
(Deeds & Rothaermel, 2003) because a relationship begins with an
initial stock of relational assets, which may include favorable be-
liefs, goodwill, trust, and commitment (Fichman &Levinthal, 1991).
Upon entering a relationship, the actors involved generally become
bound to the relationship through a justi?cation process, which
creates a certain amount of behavioral inertia (Salancik, 1997). This
viewof relationship length posits that time serves as an enervating
force, and members experience a decrease in identi?cation and an
attendance drop regarding in-role and extra-role behaviors with
time. Moreover, Mael and Ashforth (1992) asserted that identi?-
cation and affective commitment increase rapidly in the early
stages of a relationship (i.e., the honeymoon period), and this in-
crease is followed by a period of inertia.
Tsai and Bagozzi (2014) reported that the level of experience
(relationship length) of members positively moderates the rela-
tionship between behavioral intentions and contribution behav-
iors. Experienced members naturally form meaningful
interpersonal relationships with other community members
(Carlson & Zmud, 1999) and, over time, these relationships create a
sense of moral obligation in the members to contribute to the group
(Nambisan & Baron, 2010). In other words, members identify more
with the community and contribute positively to it when they
spend a long time being members of the community.
This study adopted both perspectives and proposes that the
positive relationship between identi?cation and desirable role-
behaviors in a BC is the strongest at the beginning (honeymoon
period) of the memberecommunity relationship. As time passes,
however, members experience a period of inertia, in which the
positive relationship between identi?cation and community be-
haviors may diminish. Members who overcome this period of
inertia experience a rejuvenation of a positive relationship with the
group because they become experienced members with a sense of
moral obligation to the community, which they express through in-
kind contributions (Nambisan & Baron, 2010; Tsai & Pai, 2013).
Accordingly, H8: The length of the relationship u-shaped moder-
ating effect exerts a moderating effect on the association between
identi?cation and role behaviors in a BC; speci?cally, the moder-
ating effect of membership during the honeymoon and rejuvena-
tion periods is positive, whereas that during the inertia period is
negative.
2.5. Group norms as a control variable
The internalization mechanism of the social-in?uence model
proposed by Dholakia et al. (2004) was operationalized through
group norms, which refer to the adoption of common self-guides
for meeting idealized goals shared with others because they are
considered to coincide with personal goals. The internalization
mechanism was thus de?ned as an individual member's under-
standing of and commitment to a set of goals, values, beliefs, and
conventions shared with other members. In addition to providing
knowledge and understanding regarding the objectives of the
community and the manner in which the community members
interact, accepting the community's group norms enables members
to consider themselves full-?edged members. The current study is
inspired by the study of Dholakia et al. (2004), who investigated
two key group-level determinants of the participation behaviors of
members in a community, namely group norms and social identity,
as a social-in?uence mechanism in their model. Thus, we used
Dholakia et al.'s (2004) group norm construct as a control variable
in the model developed in the current study to examine whether it
modi?es the original hypothesized path coef?cients from BC
identi?cation to desirable role behaviors.
3. Research methods
3.1. Data collection
The current study surveyed members of car clubs because,
among BCs, they are characterized by high member participation
and engagement (Algesheimer et al., 2005). The three-stage data
collection process was spread over a 4-month period between
October 2012 and January 2013 and involved recruiting re-
spondents from nine Taiwanese car clubs. The current study was
designed to investigate the causal sequence that occurs from the
initial perceptions of the characteristics of a BC to the development
of identi?cation with the organization, which then develops into
in-role and extra-role behaviors (as outlined in the conceptual
framework). First, we personally contacted either the president or
organizers of the nine car clubs and explained the purpose of our
research to obtain their support. The organizers of each club then
encouraged members to complete our survey, which was hosted on
a commercial survey website. Coupons were sent to the partici-
pants who completed the survey to express our appreciation for
their participation (worth approximately 3 USD). We collected data
in three phases to test the socialization dynamics in BC settings and
overcome the problem of common method variance that is expe-
rienced in most cross-sectional studies (Podsakoff, MacKenzie, Lee,
& Podsakoff, 2003).
Speci?cally, during the ?rst stage, we measured the constructs
of BC characteristics. In the second stage, we measured members'
identi?cation with the BC and the moderating construct of
normative pressure. In the ?nal stage, we collected data regarding
members' behaviors and their correlation with the length of the
relationship. Finally, we excluded surveys that were incomplete or
incorrectly completed before matching the responses among all
three surveys, obtaining 280 valid matched responses for further
analysis. The average age of the respondents was 33.24 years, and
253 (90.4%) of them were men. These respondents were generally
highly educated because approximately 49.6% of them held a col-
lege or higher degree. Regarding the duration of membership, 78
(27.9%) of the members were in their respective communities for
less than a year, 71 (25.4%) were in their respective communities for
1e2 years, 58 (20.7%) were members for 2e3 years, and 73 (26%)
were members for more than 3 years.
3.2. Measures and methods
In this study, the items used for our measurement scales were
based on empirically validated scales from previous studies. All
questionnaires were measured on a 7-point Likert scale anchored
by strongly disagree and strongly agree. Before each stage, we
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 194
pretested the questionnaire twice and conducted ?nal adjustments
to the wording and instructions. Table 1 shows a detailed summary
of the multiitem measurement scales. The questionnaire was
translated into Chinese according to standard back-translation
procedures (Brislin, 1970).
To evaluate the perceptions of the characteristics of a BC, during
the ?rst survey stage, we used four survey items to measure social
bene?ts, three items to measure BC prestige, and three items to
measure informativeness adapted from Dholakia et al. (2004); Tsai
and Bagozzi (2014); and Van Dolen, Dabholkar, and De Ruyter
(2007); respectively. The eight items used to measure the
emotional dimension were adopted from Wetzels, Odekerken-
Schroder, and Van Oppen (2009) and Mathwick, Malhotra, and
Rigdon (2001). To assess BC identi?cation, in the second stage of
the survey, we adopted four items from Tsai, Huang, and Chiu
(2012) and Zhou, Zhang, Su, and Zhou (2012). In the third stage of
the survey, we measured four items related to in-role behaviors and
four items related to extra-role behaviors adapted from
Bhattacharya and Sen (2003) and Ahearne et al. (2005); in addition,
a self-report measure of the length of the members' relationship
with their BCs (Tsai & Bagozzi, 2014) was used. In addition to the
previously identi?ed antecedents of BC identi?cation and their
effects on in-role and extra-role behaviors, group norms may affect
the behaviors of members. This construct acts as a control variable
and is shown in the framework in Fig. 2. The three-itemgroup norm
scale was adapted fromDholakia et al. (2004) and Chen, Arzu Wasti,
and Triandis (2007).
4. Results
According to the procedure developed by Anderson and Gerbing
(1988), we adopted a two-step approach to test the conceptual
model. First, we analyzed the measurement properties of the
re?ective latent constructs. Second, we tested the research hy-
potheses by using the SmartPLS program.
4.1. Measurement model validation
We calculated the composite scale reliability (CR; Chin, 1998)
and average variance extracted (AVE; Chin, 1998) to evaluate the
reliability of the measures. The results are listed inTable 2 and show
the internal consistency of all constructs. The CR of all measures
exceeded .90, and all the AVEs clearly exceeded the cutoff value of
.50. These results showed that our model has an adequate level of
internal consistency (Bagozzi & Yi, 1988).
As shown in Table 3, the square root of the AVE was higher than
the intercorrelations of the construct with the other constructs in
the model, thus indicating that the discriminant validity was
adequate (Barclay, Higgins, &Thompson, 1995; Chin, 1998; Hulland,
1999). Moreover, Appendix A illustrates that all items loaded
signi?cantly on their respective constructs, had standardized
loadings of at least of .70, and exhibited limited cross loadings. The
low magnitude of cross loadings compared with that of the factor
loadings also indicated that the discriminant validity is suf?cient
(Chin, 1998; Hulland, 1999).
4.2. Structural model evaluation
Tenenhaus, Vinzi, Chatelin, and Lauro (2005) proposed using the
geometric mean of the average community and average R
2
(for
endogenous constructs), where the goodness of ?t (GoF) is
0 & GoF & 1, as a global ?t measure for PLS path modeling. We
adopted the suggestion of Wetzels et al. (2009) to derive the GoF
criteria for small, medium, and large effect sizes of R
2
(in the
Table 1
Summary of measures.
Construct Measures Factor loadings
Social bene?ts 1. I joined XYZ to make new friends. .86
2. I participate in activities to stay in touch with other members. .91
3. Members in XYZ are supportive of my ideas. .92
4. I joined XYZ to expand my social networks. .89
Brand community prestige People generally regard your community as
1. Well-known. .87
2. Respected. .95
3. Prestigious. .93
Informativeness 1. The information provided by XYZ is useful. .88
2. The information provided by XYZ is valuable. .93
3. XYZ is a very good source of information. .91
Enjoyment 1. I ?nd participating in XYZ interesting. .95
2. I ?nd participating in XYZ enjoyable. .94
3. I ?nd participating in XYZ fun. .95
Escapism 1. I joined XYZ to relax. .91
2. I joined XYZ to escape from the pressure of everyday life. .96
3. I joined XYZ to escape from work pressure. .95
4. I joined XYZ to pass the time when I am bored. .93
Brand community identi?cation 1. I am very attached to XYZ. .94
2. Other community members and I share the same objectives. .93
3. I see myself as a part of XYZ. .94
4. I feel a sense of belonging toward XYZ. .88
In-role behavior 1. I often participate in activities conducted by XYZ. .84
2. I often interact with other members in XYZ. .87
3. I am loyal to the car brand belonging to XYZ. .87
4. I would purchase the same car accessories as the other members in XYZ. .86
Extra-role behavior 1. I often talk favorably about XYZ and my car to close friends, family members, or colleagues. .83
2. I often generate positive word of mouth about XYZ and my car. .85
3. I often recommend XYZ and my car to close friends, family members, or colleagues. .93
4. I try to recruit new members to join XYZ. .92
Group norm 1. Interacting with other members in XYZ can be considered to be a goal for me. .92
2. Participating in activities conducted by XYZ can be considered to be a goal for me. .91
3. My actions are often in?uenced by how other brand community members want me to behave. .81
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 195
equation GoF ¼
?????????????????
AVE*R
2
p
) and de?ne GoF
small
¼ .1,
GoF
medium
¼ .25, and GoF
large
¼ .36. These values may serve as
baseline values for validating the PLS model globally (Wetzels et al.,
2009). We obtained a GoF value of .67, which exceeded the cutoff
value of .36 for the large effect size of R
2
; we thus concluded that
the current model is favorable according to the aforementioned
baseline values. The GoF value of .67 indicated that the hypothe-
sized model provides a good representation of the structures that
underlie the observed data.
As illustrated in Fig. 2, the main effect of social bene?ts on BC
identi?cation (g ¼.28, p < .001) was positive and signi?cant, as was
that of BC prestige and identi?cation (g ¼.15, p < .05); these results
support H1 and H2, respectively. We thus concluded that the social
dimension of BC characteristics exhibits positive and signi?cant
main relationships with BC identi?cation. Moreover, the positive
effect of the cognitive dimension of informativeness on BC identi-
?cation supported H3 at a signi?cance of .05 (g ¼ .08, p < .05).
Regarding the antecedents of the emotional dimensions of BC
characteristics, signi?cant positive relationships were observed
between entertainment and enjoyment and identi?cation (g ¼ .31,
p < .001) as well as between escapism and identi?cation (g ¼ .20,
p < .001). These results supported H4 and H5, respectively.
Finally, after we controlled the group normconstruct, the results
showed a signi?cantly positive relationship between identi?cation
and the consequent construct of in-role behaviors in BCs (b ¼ .69,
p < .001); these results supported H6. Likewise, identi?cation
exerted a positive and signi?cant effect on extra-role behaviors in
BCs (b ¼ .52, p < .001), supporting H7. The values of the variance
(R
2
) in BC identi?cation, in-role behaviors in BCs, and extra-role
behaviors in BC explained by the model were .70, .59, and .39,
respectively.
4.3. Moderating in?uence of relationship length
To test the hypotheses regarding the moderating effects that the
length of the relationship exerts on identi?cation and role behav-
iors in BCs, we performed two analyses to con?rm the u-shaped
curvilinear moderating effect. The ?rst was a multigroup analysis,
and the second was a hierarchical regression analysis presented in
Appendix B, as suggested by Aiken and West (1991) and Jaccard and
Turrisi (2003).
We categorized the respondents into three groups to test the u-
shaped moderating effect of the length of the relationship on BC
identi?cation and desirable role behaviors. The thresholds used to
cluster the membership duration of participants were determined
according to (a) in-depth interviews with the presidents and or-
ganizers of nine car clubs; (b) the academic support presented in
our hypothesis development section; and (c) the percentage dis-
tribution of the membership tenure of the respondents in our data
set.
Fig. 2. Structural model results. Path coef?cients: ***p < .001; **p < .05; *p < .1.
Table 2
Evaluation of the internal consistency of the measurement model.
Constructs AVE CR R square Cronbach's alpha
Prestige .84 .94 .90
Social bene?t .80 .94 .92
Informativeness .82 .93 .89
Entertainment .89 .96 .94
Escapism .88 .97 .95
Identi?cation .85 .96 .70 .94
In-role behavior .74 .92 .57 .88
Extra-role behavior .78 .93 .36 .90
Group norm .77 .91 .85
Note: CR ¼ composite reliability; AVE ¼ average variance extracted.
Table 3
Evaluation of the discriminant validity of the measurement model.
1 2 3 4 5 6 7 8
1. Entertainment .95
2. Escapism .57 .94
3. Extra-role behavior .55 .53 .88
4. Identi?cation .77 .63 .60 .92
5. In-role behavior .68 .68 .65 .76 .86
6. Informativeness .47 .22 .32 .44 .30 .90
7. Prestige .67 .44 .49 .63 .50 .39 .92
8. Social bene?t .73 .60 .47 .74 .74 .41 .57 .89
Note: Diagonal numbers in bold are the square roots of the AVE score for each
construct.
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 196
Our sample comprised members of one of the six car clubs
centered on three car brands (Mitsubishi, Ford, and Nissan). The in-
depth interviews with the organizers and presidents of the six car
BCs provided a clearer understanding of members' participation and
relationship dynamics in the car clubs. Moreover, academic thresh-
olds for de?ning newcomers have been de?ned. For example,
Ashforth and Saks (1996) de?ned newcomers as people who had
started a relationshipwith a neworganizationwithinthe previous 10
months. Morrison (2002) used 9 months as a threshold. However,
without any speci?c consensus on the de?nition of a newcomer,
especiallyinthe BC(car club) context, we interviewedthe community
organizers and determined that a reasonable duration for a
newcomer status in a car club is 6 months to 1 year, which is
consistent withthedescriptionof the honeymoonrelationshipperiod
provided in the hypothesis development section. According to the
general consensus determined through the in-depth interviews,
some members reduce their attendance frequency and show
decreased enthusiasm regarding participation in activities after
joining a car club for approximately 2e3 years; this phenomenon is
consistent with the period of inertia. Some members may leave after
joining a car club for 3e4 years, indicating the declining phase of
relationship lifecycle. However, members who overcome this inertia
periodbecome veterans andsenior members. All organizers of the six
clubs were senior members with more than 3 years of tenure. Finally,
in our dataset, 27.9% of the respondents' memberships were shorter
than 1 year, 25.4% were members for 1e2 years, 20.7% had 2e3 year
tenures, and 26% were members for more than 3 years.
Thus, according to the presented discussion regarding the in-
depth interviews with the organizers of the car clubs, literature
review of previous studies, and statistical distribution of the
membership tenure of our respondents, we categorized the rela-
tionship length of the respondents into three groups: (a) Tenure
shorter than 1 year, corresponding to the members' honeymoon
period in the BC; (b) tenure of 1e3 years, re?ecting the inertia
period in membership to the BC; and (c) tenure of more than 3
years, re?ecting the rejuvenation period in membership to the BC.
After we separated the groups according to the length of the
relationship with the BC, we established separate structural models
on the basis of three subsamples and conducted moderation tests to
identify the differences in the respective coef?cients of the hy-
pothesized paths. As shown in Table 4, the relationship length
exerted a u-shaped effect on the relationship between BC identi?-
cation and in-role behaviors. The change of two path coef?cients
was signi?cant at p < .05 (z ¼ 2.22), indicating that, from the hon-
eymoon period to the inertia period, the positive relationship be-
tween identi?cation and in-role behaviors weakened signi?cantly.
Subsequently, from the inertia period to the rejuvenation period,
members' positive relationship between identi?cation and in-role
behaviors signi?cantly strengthened at a level of p < .01 (z ¼À1.94).
The relationship length exerted a u-shaped curvilinear moder-
ating effect on the positive relationship between identi?cation and
extra-role behaviors. Members who joined the BC for 1 year or less
exhibited the highest path coef?cients regarding identi?cation and
both in-role and extra-role behaviors. As members passed through
the inertia stages (i.e., 1e3 years of membership), the relationship
length negatively in?uenced the positive relationship between
identi?cation and role behaviors. Beyond a certain length of tenure,
the positive effect of BC identi?cation on role behaviors increased
with relationship longevity (i.e., over 3 years). The results of our
multigroup analysis indicated that relationship length moderates
the positive effect of BC identi?cation on both in-role and extra-role
behaviors in a curvilinear manner, thus supporting H8.
5. General discussion
The current research is an extension of previous studies on BCs.
This study proposed and tested a moderating-mediating model of
the in-role and extra-role behaviors of members of BCs. The ?nd-
ings show that social, emotional, and cognitive dimensions of BCs
signi?cantly in?uence role behaviors through members' identi?-
cation with the brand. Both emotional and social characteristics are
particularly crucial to this process. This study also shows that
relationship length exerts a u-shaped curvilinear moderating effect
on the positive relationship between identi?cation and desirable
role behaviors in BCs. In this section, each of these results and their
implications are discussed, and some directions for further research
are suggested.
First, this study extended the current understanding of the an-
tecedents of members' identi?cation with a BC. We identi?ed three
dimensions of BC characteristics (social, cognitive, and emotional)
that in?uence the behaviors of members through their identi?ca-
tion with the community. Among the examined factors, social
bene?ts, enjoyment, andescapismplayedrelativelymoreimportant
roles in triggering members' identi?cationwitha BC comparedwith
communityprestige andinformativeness. These results impliedthat
participants who freely join and participate in a BC can derive a
sense of ful?llment in their pursuit of happiness (hedonic charac-
teristics) and a feeling of belonging (social characteristics), which
cause them to develop a feeling of identi?cation with the commu-
nity. Both Cheung and Lee (2009) and Tsai and Pai (2013) reported
that hedonically rewarding experiences might contribute to an
understanding of the perspectives of members regarding commu-
nity engagement. The current study con?rmed the merits of pur-
suing this line of inquiry and clari?ed an underlying psychological
mechanism in consumers, that is, af?liating with a BC is a possible
avenue for developing a ful?lling social identity. Furthermore, ac-
cording to the asymmetric roles of organizational characteristics in
constructing social identities indicated by our empirical results, we
infer that such emotional and social characteristics enable in-
dividuals to de?ne themselves as part of an in-group and to cate-
gorize themselves socially according to the characteristics of the BC.
Second, we examined the effect of BC identi?cation on both in-
role and extra-role behaviors. We empirically validated that
members identify with communities and that this identi?cation
Table 4
Moderating effect of relationship length: Multigroup analysis.
Path coef?cients Standard error T-statistics
Honeymoon period Within-one-year group (Subsample N ¼ 78)
H6: BC identi?cation e> In-role .83 .06 16.31
H7: BC identi?cation e> Extra-role .69 .09 7.44
Inertia period One- to-three-year group (Subsample N ¼ 129)
H6: BC identi?cation e> In-role .67Y .04 16.56
H7: BC identi?cation e> Extra-role .55Y .08 7.20
Rejuvenation period Over three-years group (Subsample N ¼ 73)
H6: BC identi?cation e> In-role .81[ .06 13.54
H7: BC identi?cation e> Extra-role .60[ .11 5.48
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 197
has strong and positive effects on in-role behaviors; this is
consistent with the ?ndings of Cheung and Lee (2009), Koh et al.
(2007), and Wiertz and de Ruyter (2007). In addition to con-
ducting in-role behaviors regarding community participation and
brand patronage, identi?ed members performextra-role behaviors,
such as providing positive word of mouth and recommendations
and recruiting newmembers; this is consistent with the ?ndings of
Tsai and Pai (2013) and Tsai and Bagozzi (2014). Speci?cally, we
demonstrated that consumers who identify more strongly with a
BC tend to recommend both the BC and brand more often.
Third, we showed that members' relationship length exerts a u-
shaped curvilinear moderating effect on the positive relationship
between identi?cation and role behaviors. This ?nding is crucial
and novel in the organization literature because it extends and
combines two theoretical perspectives, those of Mael and Ashforth
(1992) and Tsai and Bagozzi (2014), regarding the effect of rela-
tionship length. The empirical results indicated that members
experience a honeymoon period that occurs less than 1 year into
the relationship, and this period is followed by a period of inertia
1e3 years into the relationship. During the period of inertia, the
positive relationship between identi?cation and role behaviors
weakens; this is consistent with the ?ndings of Bhattacharya et al.
(1995) and Deeds and Rothaermel (2003).
When members continue to participate in a community (over 3
years into the relationship), they become experienced and integral
parts of small friendship groups (Tsai & Bagozzi, 2014; Tsai & Pai,
2013). These meaningful interpersonal relationships create a
sense of moral obligation to participate in and support the BC. The
results obtained in the current study showed that members in this
group (i.e., those whose membership in the BC exceeded 3 years)
exhibited an increasingly positive relationship between BC identi-
?cation and extra-role behaviors; these results are consistent with
the ?ndings of Carlson and Zmud (1999) and Nambisan and Baron
(2010) and respond to suggestions by Chiu, Hsu, and Wang (2006),
Tsai and Pai (2013), Wasko and Faraj (2005), and Wiertz and de
Ruyter (2007) to model more precisely the intricacies associated
with identi?cation and BC participation. The novel ?nding of a u-
shaped curvilinear moderating mechanism associated with rela-
tionship length is a crucial step toward developing a more
comprehensive explanation of the relationship between members'
identi?cation and role behaviors in a BC.
6. Limitations and future research
When interpreting the ?ndings, we recognized that our study has
several limitations. First, it is unclear whether our ?ndings are appli-
cable to all types of BCs. We studied the members of car clubs, and
different results may be obtained when other types of communities
are sampled. Second, we used three groups of relationship lengths to
analyze the moderating effects described in H8 according to in-depth
interviews with the presidents of six car clubs, the literature review,
and the percentage distribution of members' tenure in our collected
sample. However, for different product categories, the thresholds of
the relationship life cycle may vary. For example, the new iPhone 6
may require a shorter life cycle to reach its maturity phase than does
the Apple Watch or Google Glass. Similarly, car products may have a
longer consumer relationship length than information technology
(3C) products do. Thus, to address these limitations and derive richer
insights, we suggest conducting additional studies on different types
of BCs concerning various products and industries.
Third, according to Cheema and Kaikati (2010), perceived
uniqueness may prevent consumers fromdelivering word of mouth.
Consumers who develop a strong brand identity may consider
themselves unique and may be reluctant to perform extra-role be-
haviors. We suggest that this concept be introduced in future studies
to examine the following question: When an individual plays dual
roles (i.e., as a member in a BC and a consumer who owns a brand
product), does he or she derive pride in belonging to a BC and
owning a majority-endorsed brand product or in seeking a minority-
endorsed brand or product to differentiate himself or herself from
other members in the group? Future research may attempt to
determine whether and how these factors in?uence individuals'
preferences for a conformity or uniqueness tendency and how these
tendencies in?uence desirable outcomes such as prosocial citizen-
ship behaviors in a BC, brand loyalty, positive word of mouth toward
a target brand and BC, and volitional recruitment of new members.
Acknowledgments
We are grateful to Professor Hsien-Tung Tsai at National Taipei
University for his previous contribution to this research. Thanks for
Professor Tsai's valuable opinions on the development of research
design and for his discussion which helped us developing the ideas
put forward here. We also appreciate Professor Tsai's gracious com-
ments onthe draft of this manuscript togreatlyenhance the qualityof
this paper. It would not have been possible without his help.
Con?ict of Interest
All contributing authors declare no con?icts of interest.
Appendix A
Table 5
Cross-loading analysis of discriminant validity.
Construct item Identi?cation Entertainment Escapism Extra-role In-role Informativeness Prestige Social bene?t
BC_IDENT_1 .94 .72 .55 .55 .59 .43 .58 .66
BC_IDENT_2 .93 .72 .58 .55 .56 .48 .62 .66
BC_IDENT_3 .94 .67 .57 .57 .59 .38 .59 .69
BC_IDENT_4 .88 .72 .61 .55 .64 .35 .55 .73
BC_PRESTIGE_1 .49 .57 .33 .50 .42 .39 .87 .42
BC_PRESTIGE_2 .63 .64 .45 .45 .44 .35 .95 .55
BC_PRESTIGE_3 .61 .62 .42 .41 .40 .34 .93 .57
ENJOYMENT_1 .72 .95 .52 .53 .55 .45 .63 .69
ENJOYMENT_2 .72 .94 .57 .52 .58 .42 .64 .69
ENJOYMENT_3 .73 .95 .53 .51 .54 .45 .62 .68
ESCAP2_1 .55 .50 .91 .46 .56 .16 .36 .53
ESCAP2_2 .61 .54 .96 .49 .56 .20 .43 .57
ESCAP2_3 .58 .54 .95 .50 .51 .20 .45 .53
ESCAP2_4 .61 .57 .93 .52 .56 .26 .41 .62
EX_ROLE_1 .50 .45 .50 .82 .41 .30 .36 .41
EX_ROLE_2 .50 .44 .41 .84 .39 .27 .41 .34
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 198
Appendix B
To test the u-shaped curvilinear moderating effect further, we
conducted another moderated hierarchical regression analysis. In
Step 1, we incorporated the main effects of BC identi?cation and
relationship length into the model. In Step 2, we added linear-by-
linear interaction terms between relationship length and identi?-
cation. In Step 3, we included quadratic-by-linear interaction terms
between relationship length squared and identi?cation in the
regression model. Step 3 involved testing the hypothesis that
members' relationship length moderates the relationship between
identi?cation and extra-role behaviors in a u-shaped curvilinear
manner. Tables 6 and 7 show the results of the regression analysis.
Adding the linear-by-linear interaction terms into Model 2 pro-
duced signi?cant results for extra-role behaviors (b ¼ À.03,
t ¼ À6.91, p < .001). In Model 3, adding the quadratic-by-linear
interaction terms produced a signi?cant quadratic-by-linear
moderating effect; that is, it con?rmed that relationship length
exerted a u-shaped curvilinear effect on identi?cation and extra-
role behavior, partially supporting H8 (b ¼ À.012, t ¼ À5.75,
p < .001).
References
Ahearne, M., Bhattacharya, C., & Gruen, T. (2005). Antecedents and consequences of
customerecompany identi?cation: expanding the role of relationship market-
ing. Journal of Applied Psychology, 90, 574e585.
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting in-
teractions. Newbury Park, CA: Sage.
Algesheimer, R., Dholakia, U. M., & Herrmann, A. (2005). The social in?uence of
brand community: evidence from European car clubs. Journal of Marketing,
69(3), 19e34.
Table 5 (continued)
Construct item Identi?cation Entertainment Escapism Extra-role In-role Informativeness Prestige Social bene?t
EX_ROLE_3 .56 .51 .50 .93 .50 .24 .48 .44
EX_ROLE_4 .56 .53 .45 .92 .52 .32 .48 .46
INFORMATION_1 .38 .39 .17 .24 .25 .88 .35 .35
INFORMATION_2 .34 .37 .14 .23 .31 .92 .34 .33
INFORMATION_3 .46 .49 .26 .37 .40 .90 .36 .42
IN_ROLE_1 .69 .63 .55 .46 .76 .26 .48 .70
IN_ROLE_2 .41 .39 .42 .40 .83 .34 .29 .41
IN_ROLE_3 .41 .36 .42 .35 .82 .28 .30 .39
IN_ROLE_4 .35 .33 .35 .36 .72 .26 .25 .32
SOCIAL_BEN_1 .62 .59 .51 .46 .48 .37 .51 .86
SOCIAL_BEN_2 .66 .64 .54 .39 .58 .32 .47 .91
SOCIAL_BEN_3 .65 .63 .57 .37 .62 .32 .54 .92
SOCIAL_BEN_4 .71 .74 .54 .46 .60 .45 .51 .89
Note: The factor loadings are in bold, while the other numbers relate to the cross loadings.
Table 6
Hierarchical regression analysis.
Model Unstandardized coef?cients Standardized coef?cients T Signi?cance
Beta Standard error Beta
1 Identi?cation .99 .01 .99 73.78 .00
Relationship length .01 .03 .01 .47 .64
2 Identi?cation .99 .01 .99 79.96 .00
Relationship length .78 .11 .42 6.87 .00
Length*Identi?cation ¡.03 .00 ¡.42 ¡6.91 .00
3 Identi?cation 1.01 .02 1.01 54.74 .00
Relationship length 2.13 .26 1.15 8.04 .00
Length*Identi?cation À.09 .01 À1.23 À8.11 .00
(Relationship length)
2
À.29 .05 À.97 À5.49 .00
Length
2
*Identi?cation .01 .00 1.04 5.75 .00
Note: Dependent variable ¼ extra-role behavior.
As shown in Table 7, Model 2, which included the interaction variable of identi?cation*relationship length, increased the explained variance by .3% (DF ¼ 47.81, p < .001). To
Model 3, the squared term for relationship length and its interaction term with identi?cation were added, contributing .2% of the explained variance (DF ¼ 16.75, p < .001).
Table 7
Results of regression analysis.
Model R R-square Adjusted R-square Changes of R-square Changes of F-statistic Signi?cance of F-value changes
1 .99 .98 .98 .98 6475.69 .00
2 .99 .98 .98 .00 47.81 .00
3 .99 .98 .98 .00 16.75 .00
Note: Predictor variables in model 1: relationship length, identi?cation.
Predictor variables in model 2: relationship length, identi?cation, relationship length*identi?cation. Predictor variables in model 3: relationship length, identi?cation,
relationship length*identi?cation, squared relationship length, squared relationship length*identi?cation.
Note: Dependent variable ¼ extra-role behavior.
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 199
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: a
review and recommended two-step approach. Psychological Bulletin, 103,
411e423.
Ashforth, B. K., & Saks, A. M. (1996). Socialization tactics: longitudinal effects on
newcomer adjustment. Academy of Management Journal, 39, 149e178.
Bagozzi, R. P., & Dholakia, U. M. (2006). Antecedents and purchase consequences of
customer participation in small group brand communities. International Journal
of Research in Marketing, 23, 45e61.
Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models.
Journal of the Academy of Marketing Science, 16(1), 74e94.
Barclay, D., Higgins, C., & Thompson, R. (1995). The partial least squares (PLS)
approach to causal modeling: personal computer adoption and use as an
illustration. Technology Studies, 2, 285e309.
Baumeister, R. F. (1998). The self. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.),
Handbook of social psychology (pp. 680e740). New York, NY: McGraw-Hill.
Bergami, M., & Bagozzi, R. P. (2000). Self-categorization, affective commitment and
group self-esteem as distinct aspects of social identity in the organization.
British Journal of Social Psychology, 39, 555e577.
Bhattacharya, C. B., Rao, H., & Glynn, M. A. (1995). Understanding the bond of
identi?cation: an investigation of its correlates among art museum members.
Journal of Marketing, 59, 46e57.
Bhattacharya, C. B., & Sen, S. (2003). Consumer-company identi?cation: a frame-
work for understanding consumers' relationships with companies. Journal of
Marketing, 67, 76e88.
Borman, W. C., & Motowidlo, S. J. (1993). Expanding the criterion domain to include
elements of contextual performance. In N. Schmitt, & W. C. Borman (Eds.),
Personnel selection in organizations (pp. 71e98). San Francisco, CA: Jossey-Bass.
Brislin, R. W. (1970). Back-translation for cross-cultural research. Journal of Cross-
Cultural Psychology, 1, 185e216.
Carlson, J. R., & Zmud, R. W. (1999). Channel expansion theory and the experiential
nature of media richness perceptions. Academy of Management Journal, 42,
153e170.
Cheema, A., & Kaikati, A. M. (2010). The effect of need for uniqueness on word of
mouth. Journal of Marketing Research, 47, 553e563.
Chen, X. P., Arzu Wasti, S., & Triandis, H. C. (2007). When does group norm or group
identity predict cooperation in a public goods dilemma? The moderating effects
of idiocentrism and allocentrism. International Journal of Intercultural Relations,
31, 259e276.
Cheung, C. M., & Lee, M. K. (2009). Understanding the sustainability of a virtual
community: model development and empirical test. Journal of Information
Science, 35, 279e298.
Chin, W. W. (1998). The partial least squares approach to structural equation
modeling. In G. A. Marcoulides (Ed.), Modern methods for business research (pp.
295e336). Mahwah, NJ: Lawrence Erlbaum Associates.
Chiu, C. M., Hsu, M. H., & Wang, E. T. (2006). Understanding knowledge sharing in
virtual communities: an integration of social capital and social cognitive the-
ories. Decision Support Systems, 42, 1872e1888.
Deeds, D. L., & Rothaermel, F. T. (2003). Honeymoons and liabilities: the relationship
between age and performance in research and development alliances. Journal of
Product Innovation Management, 20, 468e484.
Dholakia, U. M., Bagozzi, R. P., & Pearo, L. K. (2004). A social in?uence model of
consumer participation in network-and small-group-based virtual commu-
nities. International Journal of Research in Marketing, 21, 241e263.
Dutton, J. E., Dukerich, J. M., & Harquail, C. V. (1994). Organizational images and
member identi?cation. Administrative Science Quarterly, 39, 239e263.
Fichman, M., & Levinthal, D. A. (1991). Honeymoons and the liability of adolescence:
a new perspective on duration dependence in social and organizational re-
lationships. Academy of Management Review, 16, 442e468.
Gruen, T. W., Summers, J. O., & Acito, F. (2000). Relationship marketing activities,
commitment, and membership behaviors in professional associations. Journal of
Marketing, 64(3), 34e49.
Hausman, A. V., & Siekpe, J. S. (2009). The effect of web interface features on
consumer online purchase intentions. Journal of Business Research, 62, 5e13.
Henry, P. C., & Caldwell, M. (2006). Self-empowerment and consumption: consumer
remedies for prolonged stigmatization. European Journal of Marketing, 40,
1031e1048.
Hulland, J. (1999). Use of partial least squares (PLS) in strategic management
research: a review of four recent studies. Strategic Management Journal, 20,
195e204.
Jaccard, J., & Turrisi, R. (2003). Interaction effects in multiple regression (2nd ed.).
Thousand Oaks, CA: Sage.
Koh, J., Kim, Y. G., Butler, B., & Bock, G. W. (2007). Encouraging participation in
virtual communities. Communications of the ACM, 50(2), 68e73.
Kunda, Z. (1999). Social cognition: Making sense of people. Cambridge, MA: MIT
Press.
Mael, F., & Ashforth, B. E. (1992). Alumni and their alma mater: a partial test of the
reformulated model of organizational identi?cation. Journal of Organizational
Behavior, 13, 103e123.
Mathwick, C., Malhotra, N., & Rigdon, E. (2001). Experiential value: conceptualiza-
tion, measurement and application in the catalog and Internet shopping envi-
ronment. Journal of Retailing, 77, 39e56.
McAlexander, J. H., Schouten, J. W., & Koenig, H. F. (2002). Building brand com-
munity. Journal of Marketing, 66(1), 38e54.
McKenna, K. Y., & Bargh, J. A. (1999). Causes and consequences of social interaction
on the Internet: a conceptual framework. Media Psychology, 1, 249e269.
Morrison, E. W. (1994). Role de?nitions and organizational citizenship behavior: the
importance of the employee's perspective. Academy of Management Journal, 37,
1543e1567.
Morrison, E. W. (2002). Newcomers' relationships: the role of social network ties
during socialization. Academy of Management Journal, 45, 1149e1160.
Muniz, A. M., Jr., & O'guinn, T. C. (2001). Brand community. Journal of Consumer
Research, 27, 412e432.
Nambisan, S., & Baron, R. A. (2009). Virtual customer environments: testing a model
of voluntary participation in value co-creation activities. Journal of Product
Innovation Management, 26, 388e406.
Nambisan, S., & Baron, R. A. (2010). Different roles, different strokes: organizing
virtual customer environments to promote two types of customer contribu-
tions. Organization Science, 21, 554e572.
Novak, T. P., Hoffman, D. L., & Yung, Y. F. (2000). Measuring the customer experience in
online environments: a structural modeling approach. Marketing Science, 19, 22e42.
Pai, P. Y., & Tsai, H. T. (2011). How virtual community participation in?uences
consumer loyalty intentions in online shopping contexts: an investigation of
mediating factors. Behaviour & Information Technology, 30, 603e615.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common
method bias in behavioral research: a critical review of the literature and
recommended remedies. Journal of Applied Psychology, 88, 879e903.
Rojek, C. (1993). Disney culture. Leisure Studies, 12, 121e135.
Rothaermel, F. T., & Sugiyama, S. (2001). Virtual Internet communities and com-
mercial success: individual and community-level theory grounded in the
atypical case of TimeZone.com. Journal of Management, 27, 297e312.
Salancik, G. (1997). Commitment and the control of organizational behavior and
belief. In B. M. Staw, & G. R. Salancik (Eds.), New directions in organizational
behavior (pp. 1e54). Chicago, IL: St. Clair Press.
Scott, S. G., & Lane, V. R. (2000). A stakeholder approach to organizational identity.
Academy of Management Review, 25, 43e62.
Tenenhaus, M., Vinzi, V. E., Chatelin, Y. M., & Lauro, C. (2005). PLS path modeling.
Computational Statistics & Data Analysis, 48, 159e205.
Tsai, H. T., & Bagozzi, R. P. (2014). Contribution behavior in virtual communities:
cognitive, emotional, and social in?uences. MIS Quarterly, 38, 143e167.
Tsai, H., Huang, H., & Chiu, Y. (2012). Brand community participation in Taiwan:
examining the roles of individual-, group-, and relationship-level antecedents.
Journal of Business Research, 65, 676e684.
Tsai, H. T., & Pai, P. (2013). Explaining members' proactive participation in virtual
communities. International Journal of Human-Computer Studies, 71, 475e491.
Van Dolen, W. M., Dabholkar, P. A., & De Ruyter, K. (2007). Satisfaction with online
commercial group chat: the in?uence of perceived technology attributes, chat
group characteristics, and advisor communication style. Journal of Retailing, 83,
339e358.
Wasko, M. M., & Faraj, S. (2000). It is what one does: why people participate and
help others in electronic communities of practice. Journal of Strategic Informa-
tion Systems, 9, 155e173.
Wasko, M. M., & Faraj, S. (2005). Why should I share? Examining social capital and
knowledge contribution in electronic networks of practice. MIS Quarterly, 29,
35e57.
Wetzels, M., Odekerken-Schroder, G., & Van Oppen, C. (2009). Using PLS path
modeling for assessing hierarchical construct models: guidelines and empirical
illustration. MIS Quarterly, 33, 177e195.
Wiertz, C., & de Ruyter, K. (2007). Beyond the call of duty: why customers
contribute to ?rm-hosted commercial online communities. Organization
Studies, 28, 347e376.
Wuyts, S. (2007). Extra-role behavior in buyer-supplier relationships. International
Journal of Research in Marketing, 24, 301e311.
Zhou, Z., Zhang, Q., Su, C., & Zhou, N. (2012). How do brand communities generate
brand relationships? Intermediate mechanisms. Journal of Business Research, 65,
890e895.
H.-Y. Hung, T.-L. Lin / Asia Paci?c Management Review 20 (2015) 191e200 200

doc_362063973.pdf