Social interaction and continuance intention in online auctions: A social capital perspective

Decision Support Systems 47 (2009) 466–476

Contents lists available at ScienceDirect

Decision Support Systems j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / d s s

Social interaction and continuance intention in online auctions: A social capital perspective Jyun-Cheng Wang a, Ming-Jiin Chiang b,⁎ a b

Institute of Service Science, National Tsing Hua University, Hsinchu, Taiwan, ROC Department of Information Management, National Chung Cheng University, Chiayi, Taiwan, ROC

a r t i c l e

i n f o

Article history: Received 3 September 2007 Received in revised form 17 January 2009 Accepted 15 April 2009 Available online 3 May 2009 Keywords: Online auction Social capital Continuance intention Social network

a b s t r a c t This study explores how interaction within an online auction community affects online auction actor intention to continue trading with others. Adopting a social perspective drawing on social capital theory and IS literature, this study investigates how interactions among actors contribute to the creation and advancement of social capital. The analytical results demonstrate that the influence of user interaction on continuance intention in online auctions is mediated by the creation of various dimensions of social capital at the community level. Finally, the implications of the study findings are discussed. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Online auctioning has proven to be one of the most successful and rapidly growing business models for electronic commerce. Users of online auctions range from individuals conducting online “garage sales”, to companies liquidating unwanted inventory [7], and business-to-business “hubs”. However, unlike traditional auctions, online auction transactions are limited by two forms of uncertainty: uncertainty regarding the quality or condition of the goods for sale and uncertainty regarding trader trustworthiness, owing to the ease of creating pseudonyms at will. To overcome these deficiencies, online auctions tend to rely on interaction mechanisms such as trader reputation ratings, online Q&A, and message boards to help potential buyers make purchasing decisions and enhance the overall trading experience [82]. Numerous studies have shown that reputation systems are crucial in building trust and thus encourage trading among strangers online [16,28,29,51,56]. However, recent research has pointed out a relatively incoherent relationship between reputation (which is an objective construct) and trust (which is a subjective construct) [41]. Additionally, despite unfair or dishonest practices such as inflation of reputation becoming increasingly prevalent in online auctions [7], people appear unconcerned with the risks associated with online transactions. Thus, the relationship between interaction mechanisms (such as reputation systems) and trustworthiness behaviors is unclear. For online auction management teams, continued member participation is one of the key determi⁎ Corresponding author. E-mail addresses: [email protected] (J.-C. Wang), [email protected] (M.-J. Chiang). 0167-9236/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.dss.2009.04.013

nants of virtual community success. Researchers and online practitioners face an important problem, namely why people continue to trade with strangers online given the uncertainty and the influence of interaction mechanisms (such as reputation systems) on their attitude and online behavior, such as trust/trustworthiness and continuance intention/behavior. This question has not yet been explored and remains vague and unsystematic. Previous studies have classified the motivations of continuance usage of web sites into system attributes (such as information quality, usefulness, and ease of use) [9,25,30,54] and individual attributes (such as trust, loyalty, and satisfaction) [13,14,35,64]. However, analysis based only on these two types of attributes is insufficient for online services in that it ignores the effects of interaction within online communities. To fill this gap, several studies have put emphasis on the fact that online auctions not only create a new transaction way but also organize auction participants into communities [5,7,18,24,28,51,72,77]. Meanwhile, several studies [5,41,44] highlighted that social factors (such as social interaction and social structure) can influence online auction user behavior. For example, in [41] they argued that trust is a subjective construct and reputation is “a quantity derived from the underlying social network”. Accordingly, the subjective trust of social actors can be considered “a combination of received referrals and personal experience”. That is, social factors such as “transitive trust path” can help community members derive trust. Furthermore, the largest online auction website in the world, eBay, recently launched a new social networking based application called ‘neighborhoods’ [45]. This new feature encourages users to establish personal communities by posting photos, product reviews, tips and responses. Users thus can provide a more visual and interactive experience environment than that offered by the previous

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

text interface, enhancing member usage of auction services. This news reveals that examining social factors of an online community can shed light on user online interaction behaviors. However, despite all the attention paid to this area, there have been few empirical studies examining continuance intention of web sites by adopting the perspective of social interaction in online communities. To incorporate the possible influence of interaction on behaviors online, this study adopts a social perspective by examining both the nature of online interaction and the social context of individual interactions. This perspective resonates with the following studies that draw attention to the importance of social environments. First, Lamb et al. [46] suggested that theories regarding the use of information and communication technologies (ICTs) should consider the importance of complex social environments. Lamb contended that most users of ICTs applications interact with others in multiple social contexts, and thus IS researchers should conceptualize ICTs users as social actors to “more accurately portray the complex and multiple roles that people fulfill while adopting, adapting, and using information systems.” In online auction environments, interaction between buyers and sellers can be considered an example of voluntary associations using ICTs. Such voluntary behavior can be one of the sources nurturing norms, trust, and other collective resources that are essential to community growth [50,62,69]. In examining the drivers of community-based ICTs, the social interaction perspective helps shift the focus from the individual to the community level and thus helps in considering the interplay between individuals and the community. In the light of social interaction within online communities, the following sections consider the term “social actor” to represent online auction players who may act as sellers, buyers, or both. Second, most studies of trust in online business models have tended to focus on trust as an antecedent to initial web usage or acceptance, and have devoted little attention to its effects on continued usage. Previous studies [6,7,21,35,59,64] have identified various types of trust: personality-based trust [35,83], knowledgebased trust [70], institution-based trust [35,65,83], calculative-based trust [65], cognition-based trust [35], and relational trust [65,70,83]. Furthermore, trust has been identified as a crucial influence on risky transactional relationships, such as online transactions. For example, in the study of Pavlou and Gefen [35,59], institutionalization of trust, such as market-driven reputation systems or third-party involvement, was considered a primary means of creating effective online marketplaces. Their study suggested the importance of establishing trust mechanisms in attracting traders, whether naive or experienced, who have not previously traded in a specific marketplace. While trust mechanisms are essential in attracting first time traders, it is interesting to consider why two auction sites with identical trust mechanisms may not achieve equivalent success in fostering communities. To understand why online auction actors continue to participate in online auction communities and conduct business with strangers, it is necessary to go beyond the antecedents to initial trust and examine the causes of continuance intention. This study focuses on the social context within which an online site mobilizes sense of community to improve trust relationships. This study attempts to understand how interaction within an online auction community affects the intention of the online auction actors to continue trading. This study views online auction sites as social contexts for interaction among social actors, namely traders. The process of linking social interaction and intention to continue a relationship involves numerous interrelated factors. Among possible approaches, this study draws upon social capital theory to propose and empirically examine a research model of continuance intention. Previous studies employed the concept of social capital to assess the quality of relationships resulting from social interactions within communities and how they affect the ability of actors to achieve shared goals [4,63]. Social capital is sometimes viewed as a pool of resources that should be tapped via the social ties embedded in social

467

networks. From this perspective, social ties can be conceived by community members as certifications of individual “social credentials” [50]. Furthermore, social capital is considered a proxy for the ability to mobilize resources embedded in relational networks (such as Q&A or reputation interaction) to achieve higher status [50] via “a series of non-negotiated or reciprocal exchanges” [71]. This study examines the performance of online auction as evaluated by actors in terms of continued usage intention. Specifically, this work investigates how actor interactions contribute to creating and advancing social capital, and how social capital reciprocally promotes further internet auction use. A model based on Nahapiet and Ghoshal's three dimensions of social capital [57] is employed for empirical testing to understand how social capital is fostered and resources mobilized within an online auction community through interaction among actors. This study contributes to understanding of online communities and thus helps provide a detailed explanation of how and to what extent online auction sties and actors with different levels of reputation and interactive resources influence the continued use of online auction services. The remainder of this paper is organized as follows: the first section reviews the literature on IS continuance usage and social interaction, and discusses the three dimensions of social capital in the transaction-oriented virtual community. Second, the research model is developed and empirically tested using survey and objective data, namely trading volume from Yahoo! Taiwan Auctions, the largest auction website in Taiwan. Finally, the results of model testing are presented and their implications discussed. 2. Theoretical background With growing numbers of social interactions occurring online, researchers have shifted their focus from real world to online platforms to explore the effects of social interaction on social capital [15,26,31,69,74,78]. Since some social interactions may “carry more valued resources and exercise greater power” [50], the same phenomenon may also occur for online auction traders. For example, consider a situation involving sellers A and B both selling the same product at the same price. Seller A is involved in numerous discussion and Q&A sessions, while seller B does not participate in any of such discussions. From the perspective of the interaction network, seller A may occupy a more valuable position than seller B. Furthermore, some buyers may be inclined to bid on the product of seller A rather than that of seller B. Previous studies have observed that theoretically, social interaction is the fundamental cause of social capital [50,57] in physical setting while other studies have shown that online social capital can be fostered via online interaction, increasing the effectiveness and efficiency of online community activities. However, different types of social interaction exist in online settings, including forum actors asking and answering questions, traders conducting online bids or transactions, individuals browsing the blogs of friends, and classmates and families sharing messages and symbols via MSN Messenger. This study thus first introduces the essential concept of social interaction in online auctions to identify the properties of social interaction among auction actors. 2.1. Social interaction in online auctions Some studies have defined social interaction as “ties”[78] or “interaction ties” [26,75]. Regardless of the terminology used, social interaction refers to a link established via reciprocity behavior between two actors. Most studies dealing with online social interactions have been limited to knowledge-oriented communities [26,74,78], such as discussion forums regarding specific topics or open source software forums [31]. Rather than focusing on knowledge-based communities, this study focuses on online auction

468

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

communities. This study examines social interactions against a backdrop of online auctions for several reasons. First, online auctions are a transaction-oriented community characterized by economic risks involving possible monetary losses. In such an environment the importance of trust in making trading decisions is increased. Second, since it is essential to select trustworthy trading partners, online auction communities tend to rely heavily on reputation scores and other online data in evaluating trading partners. Compared with knowledge communities, auction community members have stronger incentives to learn more regarding trader transaction history to minimize trading risks. In online auction communities, the social networks embedded in online social interactions represent community social capital [69]. Meanwhile, in this study, the social interaction of a social actor represents the foundation of the three dimensions of social capital proposed by Nahapiet and Ghoshal [57]. In the context of online auctions, a dyadic link can be established during two transaction stages, namely the bidding and transaction-commit stages. During the bidding stage, since transactions have not yet been finalized, links are established between any pair of question issuers and answer providers in the public space. After a transaction is completed (that is, during the transaction-commit stage), only the two actors (i.e. the buyer and the seller) involved in the transaction can establish links that may involve payment, delivery, and reputation rating. Based on these two stages of interaction behavior, this study suggests the existence of two potential forms of social interaction among social actors, namely bidding stage interaction (or Q&A-based interaction during the bidding stage) and transaction-commit stage interaction (or reputation-based interaction during the transaction-commit stage). The two stages of social interaction are not necessarily linked in nature. First, bidding stage interaction is designed mainly to obtain more information regarding goods, while transaction-commit stage interaction is designed to evaluate overall party behavior after

Table 1 Three dimensions of social capital: element and reference. Dimension

Element

Structural Network dimension configuration

Interaction ties

Relational Trust and dimension trustworthiness Reciprocity

Identification Cognitive Shared belief dimension Shared vision

Social engagement

Reference • “The overall pattern of connections between actors” [57] • “Network ties between actors; network configuration or morphology describing the pattern of linkages in terms of such measures as density, connectivity, and hierarchy” [57] • “The ability of individuals to make connections to others within an organization”[48] • “The connections between individuals, or the structural links created through the social interactions between individuals in a network”[78] • “The predictability of another person's actions in a given situation”[48] • “The setting of common standards of behavior that individuals are willing to abide by” [48] • “A sense of mutual reciprocity, for example, the willingness to return a favor with a favor” [48] • “The process whereby individuals see themselves as united with another person or set of individuals” [48] • “Shared representations, interpretations, and systems of meaning among parties” [57] • “Shared understanding between parties” [78] • “Like a shared code and or a shared paradigm that facilitates a common understanding of collective goals and proper ways of acting in a social system” [75] • “Shared narratives or stories that can enable individuals to make sense of their current work environment and their relative role within it.” [48] • “Interacts over time with others sharing the same practice and learns the skills, knowledge, specialized discourse, and norms of the practice.” [78]

Table 2 Relationship between different elements of social capital and knowledge sharing community.

Structural dimension Relational dimension Cognitive dimension

Wasko and Faraj [78]

Chiu et al. [26]

Centrality Commitment reciprocity Self-related expertise tenure in the field

Social interaction ties Trust norm of reciprocity identification Shared language shared vision

completing a transaction. Second, a positive comment can result from positive perceptions of repeated interactions (such as negotiation for price, paying and logistics) and can be considered the reciprocal contact during the transaction-commit stage, while Q&A can easily be created and can be considered a type of accidental interaction or contact during the bidding stage. On the one hand, an actor may be awarded a negative reputation rating during the transaction-commit stage or give up on bidding in spite of having frequently engaged in Q&A interaction during the bidding stage. On the other hand, an actor with high reputation score may have low Q&A interaction due to their goods being standard products such as books or memory cards. That is, both types of social interaction, namely Q&A-based interaction and reputation-based interaction, could be considered the two different types of interaction records of auction actors that occur during different stages of participation in an auction. 2.2. Continuance intention in online auction This study defines continuance intention as user plans to continue using already adopted online auction systems [14]. Continuance intention to usage is a conceptually distinct construct from initial adoption. The latter occurs before initial acceptance of an IT/IS, online service, or product. For example, technology acceptance models (TAM or TAM-2) [27,76] explain why users adopt new systems or online service. However, such models do not investigate reuse of systems or online services post-adoption [74]. Previous studies have pointed out that continuance intention exerts a key positive influence on the success of online communities, with effects including greater community participation in online communities [3]. Furthermore, continuance intention is central to the internet marketing/EC context. For example, attracting new customers may be five times as costly as retaining existing customers [14]. Since online auctions are a commercial context as well as a virtual community, this study suggests that understanding continuance intention of auction actors is essential to the online auction management. 2.3. Key elements of social capital in online auction Despite the widespread application of social capital theory, its succinct definition remains challenging. This lack of conceptual clarity contributes to excessive-versatility [52]. Recently numerous studies have identified three dimensions of online social capital: structural, cognitive and relational [1,57,75,78,79]. The cognitive dimension refers to resources that enable shared interpretations and meanings within a group [57,78]. Furthermore, the cognitive dimension comprises values, attitudes, beliefs, and perceptions of support affecting interdependence [42,57]. Meanwhile, the relational dimension involves social actors trusting other actors within the group and being willing to reciprocate favors or other social resources in the trading process [57,78]. The structural dimension is created when community actors communicate each other. Table 1 lists the elements involved in the three dimensions of social capital. Given the varied components of each of the three dimensions of social capital as presented in the literature, the different types of interaction media (such as Internet vs. Face-to-Face) and communities

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

Fig. 1. Research model.

(such as knowledge-sharing communities vs. transaction-oriented communities) also add to the differences in considering the composition of the three dimensions of social capital. For example, two studies on knowledge-sharing community, Chiu et al. [26] and Wasko and Faraj [78] identified different elements of social capital. Table 2 compares the elements identified between the two studies. Thus, it may be necessary to identify the key elements of social capital in transaction-oriented communities by clarifying the characteristics of online auction interactions. Transactions and interactions between two parties, such as Q&A and reputation, are recorded in auction websites following the completion of a trade. Some studies have defined social interaction as either “ties”[78] or “interaction ties” [26,75]. Regardless of the terminology used, all these studies have focused on capturing community interaction patterns, which correspond to the meaning of the structural dimension of social capital. This study considers “social interaction” a form of structural dimension, and employs positive reputation and Q&A records to measure social interaction. The relational dimension of social capital typically focuses on trust, commitment, or reciprocity [57,75]. Trust is essential in creating friendly online transaction environments [6,7,35,43,73]. Auction actors can only assess the condition of goods via pictures, descriptions, and Q&A, and also have only limited information about their transaction counterparts in general [18]. Designing an environment that provides traders with increased certainty and promotes trust therefore is important. This study thus considers “trust” to be the key element in the relational dimension of social capital in online auctions. From the community perspective, shared vision is also a key element of cognitive dimension of social capital in the study of Tsai and Ghoshal [75]. Shared vision significantly affects actor cohesion and community fostering for various community types. Royal and Rossi [66] noted that a shared vision reflects “the influence a group may have over its members by encouraging commitment to a common set of ideas” and “may lead group members to feel that they share a common future.” [66] Shared vision can be critical in creating a meaningful sense of community. Online auctions represent one type of community, and teams managing auction websites should note that sense of community is vital for achieving positive outcomes in terms of business objectives for auction websites [18,55,66]. This study thus considers “shared vision” to be the key element in the cognitive dimension of social capital. 2.4. Model and hypothesis Fig. 1 depicts the research model examined in this study by displaying the hypothesized relationships among key elements of social capital and intention to continue trading with others via an auction site, named auction continuance intention (ACI). 2.4.1. Linking shared vision and auction continuance intention Tsai and Ghoshal [75] argued that organization members with high level shared vision have enhanced perceptions regarding how to communicate with other members and minimize misunderstandings

469

during their interaction. An auction actor with a detailed understanding of transaction risk during the pre-transaction stage may have higher continuance intention for conducting transactions. In the study of Burroughs and Eby [22] of psychological sense of community in the workplace, group members with a clear, shared vision were found to be of greater team orientation and sense of community, and thus increased intention to interact with others. The job of actors in online auctions is to search for and evaluate products, negotiate and transact with other actors. Actors who perform their job with a high level of shared vision and team orientation may gain positive perceptions from the community and auction actors thus may become predisposed towards continuous action. Based on this argument, the following hypothesis is established: H1. The shared vision of actors is positively associated with their auction continuance intention. 2.4.2. Linking trust and auction continuance intention The relational dimension of social capital exists when actors have a definite affinity with the group, trust other actors within the group, and are willing to reciprocate mutual actions during trading [12,78]. As mentioned previously, this study contends that trust is one of the key elements in the relational dimension of social capital in transaction-based online communities. According to the studies of Paldam and Svendsen [58] and Pollack et al. [61], trust describes the mutual expectation (for example, members acting in mutually supportive ways, or at least not intentionally harming others) arising within a community characterized by regular cooperative behaviors and commonly-shared norms. With the formation of trust relationships between two actors, the actors concerned become more likely to improve their collaboration efficiency by facilitating coordinated actions [63]. Consequently, it is reasonable to expect that social actors with high level of trust in other actors can acquire higher continuance intention to interact with their exchange partners [75]. This study thus hypothesizes: H2. Actor's trust in others is positively associated with their auction continuance intention. 2.4.3. Linking social interaction and auction continuance intention Social interaction, a key element of the structural dimension of social capital, refers to the overall pattern of connections between actors [23]. That is, social interaction describes the impersonal configuration of linkages between people or units [57], as well as the collective behavior emerging from electronically-enabled human networks and its significance in terms of advancing human development globally [1]. When auction members communicate with one another, social interaction is fostered by the communication process. Social interaction can be described as having such measurable qualities as density, connectivity, and hierarchy [57]. Wasko et al. [78] demonstrated that social interaction is important in knowledge exchange. Actors who are central to a network and connected to large numbers of other actors are more likely to continue to contribute to collective activity [23], which is important in sustaining community interaction via electronic networks [53]. In the context of online auctions, actors with more active record of participation in Q&A (that is, those that exhibit extensive Q&A records embedded in transactions) or higher positive reputation scores represent a high level of social interaction during different stages of the transaction process. This study thus contends that an auction actor with dense connectivity (e.g. higher reputation) to other actors has a high level of social interaction, and will obtain a greater structural dimension of social capital and enjoy access to more resources. This aggregate of social interaction can facilitate resource exchange [33] and bring further value or benefits to auction actors, including increased intention to continue interacting. Meanwhile, intensive

470

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

social interaction is important in facilitating information exchange [49], in turn helping auction actors avoid excessive expectations or misunderstandings in transaction and communication. Satisfaction with interaction thus may emerge when “the extent to which their expectation is confirmed” and “expectation on which that confirmation was based” [14]. It is reasonable for satisfied actors to be more likely to have increased intention to interact with other actors. Hence, this study hypothesizes: H3a. The degree of actors' bidding stage interaction is positively associated with their auction continuance intention. H3a. The degree of actors' transaction-commit stage interaction is positively associated with their auction continuance intention. 2.4.4. Relationships among different elements of social capital Social capital has been analytically divided into the three dimensions identified above. Scholars have also suggested the existence of close interrelationships among the characteristics contained in the above three categories [57]. This study examines the interplay among these elements in the three dimensions of social capital. A strong sense of community encourages the development of shared vision [22], thus facilitating the development of “common standards of behavior,” or “common values and beliefs” [48]. Moreover, shared vision helps increase awareness of how individuals react in given situations, and thus the likelihood of trusting relationships developing among actors [48,49,57]. This study thus hypothesizes: H4. Shared vision of actors is positively associated with their trust in other actors. Second, previous studies have documented that dense social interaction and ties help social actors or group members “realize and adopt their organization languages, codes, value, and practices” and share a collective orientation conducive to the formation of a group/organization vision [49,75]. As actors interact with one another, they become more likely to develop “new sets of value or new vision” based on social interaction and mutual understanding [75]. This study thus assumes that auction actors are likely to share a common value or vision with other actors via the dense process of social interaction and communication. Consequently, this study hypothesizes: H5a. The degree of bidding stage interaction is positively associated with shared vision of online auction actors. H5b. The degree of transaction-commit stage interaction is positively associated with shared vision of online auction actors. Numerous studies have suggested that interpersonal affect, trust, and trusting relationships generally result from strong, symmetrical interaction ties [36,38,57,75]. On the one hand, frequent social interaction leads to actors sharing more information or common perspectives with other actors and creating trusting relationships; on the other hand, as Tsai and Ghoshal noted [75], “an actor occupying a central location in a social interaction network is likely to be perceived as trustworthy by other actors in the network.” This study thus expected an auction actor with high social interaction to be likely to develop good judgment regarding the behavior of other actors and to enhance their trust in other actors based on their experience of interaction. Moreover, for auction actors interacting with others via an auction site where users generally use pseudonyms, reputation helps decide whether to engage in a transaction. The availability of transaction histories and reputation can potentially encourage trustworthy behavior and discourage negative actions. An auction actor with a positive reputation, manifesting in the form of social interaction, or who is an active discussion participant tends to occupy

a central position in a collective network. This is more likely to induce trust from others, thus initiating new transaction opportunities. This study thus hypothesizes the following: H6a. The degree of bidding stage interaction is positively associated with trust of online auction actors. H6a. The degree of transaction-commit stage interaction is positively associated with trust of online auction actors. 3. Research design 3.1. Instrument construction The measures used to operationalize the model constructs are primarily adapted from previous related studies, with minor changes in wording to ensure they fit the target context. Four constructs are measured: social interaction, shared vision, trust in other actors, and auction continuance intention. The first construct, social interaction, originates from actor interaction ties [78]. This study treated the two types of interaction ties, Q&A-based interaction and Reputation-based interaction, as measurable indices of two stages of social interaction, namely the bidding stage and transaction commit stages, respectively. Interactions among traders in the Q&A forums for specific products reflect their shared common interest in the product, similar to individuals browsing or discussing a topic in a virtual community. This study uses interaction data from a Q&A forum on a product to reflect the social interaction experiences of traders involved in the bidding stage. All Q&A data revealed in actor historical transaction records were aggregated as the value of Q&A-based interaction. The product delivery and payment processes follow the completion of a transaction. All these processes are personal interactions, as opposed to the open Q&A discussions, and their experiences are usually summarized in reputation ratings. This study uses aggregate positive reputation scores to represent positive interaction experiences which occur during the transaction-commit stage. Negative and neutral ratings are not counted toward Reputation-based interaction, because they represent unacceptable behaviors and imply a negative impact on the relationship. Both Q&A-based interaction and Reputation-based interaction are transformed using log transformation to fit the normality assumption of SEM analysis. From a network perspective, Reputation-based interaction represents the aggregate transaction relationships of an ego, which can be considered as the Degree Centrality measures of the ego network of a seller. The second construct, shared vision, contains multiple-item scales drawn from the social capital literature [20,26,47,57,75]. Previous studies suggest that a shared vision helps communities or group actors embody their collective beliefs, actions and aspirations [26,57,75] and develop shared images of the future they seek to create [34]. Different activities and community goals between virtual communities lead to differences in operationalization. The three elements, collective beliefs, actions, and aspirations, have been used by [26] to capture concepts of shared vision in other type of virtual community. In an auction community, attracting more actors to the community and encouraging them to actively engage in buying and selling can enhance such benefits as increased market liquidity [60]. Thus, a popular market and the existence of active actors are an important shared image of the future and a common sense of the community purpose in an auction community. With such a popular market, auction actors will be able to transact with various actors in an efficient way and enjoy the positive network externalities (e.g. lower search cost or more choices) [60]. A three-item construct is identified, which consists of: (1) “I see the auction website as popular” (2) “The auction website always comes to mind when I want to purchase something” and (3) “The popularity of the auction website exceeds

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

471

Table 3 Summary of the survey items used in this study. Constructs

Items

Bidding stage interaction Transaction-commit stage interaction Shared vision (SV) SV1 SV2

Means Q&A-based interaction Reputation-based interaction

I see the auction website as popular. The auction website always comes to mind when I want to purchase something. SV3 The popularity of the auction website exceeds my expectations. Trust in actors (Trust) Trust1 I believe that my trading partner is good and honest. Trust 2 My trading partner treated me with the respect I expected during the trading process. Trust 3 My trading partner meets my expectations in terms of honesty. Auction continuance ACI1 I intend to continue interacting with actors of auction websites. intention (ACI) ACI2 I intend to continue interacting with actors in this auction website rather than seeking out actors of other sites. ACI3 If I could, I would like to discontinue my interaction with the actors of this auction website.

my expectations.” These three elements are formulated to measure collective aspirations, actions, and beliefs of auction community actors. Table 3 shows the three measures developed to assess the shared vision of an auction community. Trust in actors was assessed by items adapted from [47,61] to reflect the beliefs of actors about honesty of trading partners, actual behaviors during the trading process, and the extent to which the actor expectations of are met. Table 3 shows these items. Finally, Table 3 shows the three items for measuring auction continuance intention based on pre-validated measures in the Expectation– Confirmation Model (ECM) developed by Bhattacherjee [14]. Respondents were asked to record their responses using a fivepoint Likert scale ranging from strongly disagree (=1) to strongly agree (=5). To ensure the balance and randomness of the questionnaire items, all items were randomly arranged to reduce the potential for floor effect that induces monotonous responses to the items used for construct measurement. Table 3 summarizes the survey items and their descriptive statistics.

Std. deviation References

246.07 844.88 592.54 3378.11

Ahuja et al. [2] and Wasko et al. [78]

4.45 4.24

0.59 0.66

Extended from Landry's beliefs. [47] and Bryant and Norris's social engagement and support [20]

4.25 3.93 4.10

0.67 0.64 0.65

Extended from Pollack and Knesebeck's civic trust [61]

4.00 4.13 4.10

0.62 0.57 0.59

4.19

0.71

Extended from Bhattacherjee's IS continuance intention [14]

contends that the online auction community is a hidden population and thus standard sampling or statistical methods are of limited use in accurately estimating its characteristics [67]. This study selected ten seeds (namely the first auction actors included in the sample) from different product categories in the website (see Table 4) and used “snowball sampling” to gather 1000 samples of auction actors. Table 4 lists the number of samples extracted from each product category. A pilot test was conducted involving 32 graduate students with experience of using online auction websites. Participants were asked to provide feedback regarding the length of the survey and the relevance and wording of the questionnaire items. Each auction actor selected via snowball sampling received an e-mail soliciting their participation. Respondents were directed to complete an online survey. This study adopted an online survey method owing to its fit with the context of online auction, as well as other advantages including lower costs, faster responses, and geographically unrestricted sample [78]. Additionally, data for both interaction ties were obtained directly from the Yahoo! Taiwan Auction website using a self-developed web crawler program.

3.2. Sampling and subjects The study sample was taken from the largest online auction website in Taiwan, Yahoo! Taiwan Auction. As mentioned earlier, social actors in online communities generally represent themselves with pseudonyms. Auction actors have limited ability to understand large quantities of information (such as sex, education, or age) about other actors. Most actors only make limited information available, for example reputation, Q&A records, and joining date. This study thus

Table 5 Respondent demographics. Measure

Items

Gender

Female Male ≧18 years old 19–24 years old 25–30 years old 31–36 years old 37–42 years old 43–47 years old ≦48 years old Student Public servant Teacher Businessman Worker Free worker Service correlation Others High school or below College (2 years) University Graduate school or above b6 months 6 months–1 year 1 year–2 years 2 years–3 years Over 3 years

Age

Table 4 Subject sampling. Product category

Select amount

%

Computers and software Cameras Cell phone Jewelry and watches Women's clothing and accessories Men's clothing and accessories Toys and video games Sports and outdoors Beauty and health Music and movies Home and garden Antiques and art Idols Books and comics Automotive Mother and baby Total amount

27 12 29 90 284 38 61 79 53 53 81 57 7 78 7 44 1000

2.70 1.20 2.90 9.00 28.40 3.80 6.10 7.90 5.30 5.30 8.10 5.70 0.70 7.80 0.70 4.40 100

Occupation

Education

Member history

38.60% 61.40% 1.50% 34.70% 35.60% 17.30% 7.40% 1.50% 2.00% 38.10% 3.50% 2.50% 18.30% 4.00% 15.80% 14.90% 3.00% 17.80% 22.30% 42.10% 17.80% 23.80% 15.30% 30.70% 13.90% 16.30%

(78/202) (124/202) (3/202) (70/202) (72/202) (35/202) (15/202) (3/202) (4/202) (77/202) (7/202) (5/202) (37/202) (8/202) (32/202) (30/202) (6/202) (36/202) (45/202) (85/202) (36/202) (48/202) (31/202) (62/202) (28/202) (33/202)

472

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

Table 6 Goodness-of-fit of the measurement model.

Table 8 Goodness-of-fit of the structural model.

Fit indices

Guidelines

Result (all)

Fit Indices

Guidelines

Result

χ2 (Chi-square) χ2 / df Goodness of fit index (GFI) Adjusted for degrees of freedom (AGFI) Normed fit index (NFI ) RMSEA CFI

Small is better b3 N 0.9 N 0.9 N 0.9 b .08 N .90

34.684 1.445 (df = 24) 0.962 0.929 0.94 0.047 0.98

χ2 (Chi-square) χ2 / df Goodness of fit index (GFI) Adjusted for degrees of freedom (AGFI) Normed fit index (NFI ) Root mean square error of approximation (RMSEA) Comparative fit index (CFI)

Small is better b3 N0.9 N0.9 N0.9 0.05–0.08 N0.9

52.198 1.135 (df = 46) 0.959 0.931 0.94 0.026 0.992

Following a single round of data collection and after excluding incomplete responses, a response rate of approximately 20.2% is achieved. The low response rate resulted partly from the use of e-mail for solicitation, which inevitably meant that some solicitations were consigned to spam or junk mail folders. Table 5 lists respondent demographic information, and reveals that the majority of respondents are ‘Male’, ‘19–30’ years old, and ‘University students’. It is not surprising that actors participating in the online auction are young men or university students with limited income and ample leisure time. As mentioned above, it is hard to verify the response bias due to the limitation of sampling from hidden population [67]. Some limitations on the sampling are outlined below in Section 5.4.

4. Results 4.1. Evaluating the measurement model As described above, this study estimated social interaction by measuring Q&A and reputation records of actors. Only three multipleitems constructs, namely auction continuance intention, shared vision, and trust in actors, were subjected to confirmatory factor analyses (AMOS 5 software). Table 6 lists the goodness-of-fit of the measurement model and Table 7 lists the scale properties (standardized item loading and t value). In line with the suggestion of O'LearyKelly and Vokurka (1998), and Hair et al. (1998), two common statistic tests were applied for assessing the unidimensionality of a measure: the overall model fit (e.g., using χ2/df statistics) and the significance of individual item loading (λ). Most of fit indices in Table 6 indicate adequate model fit, and all estimated standard loadings in Table 7 are statistically significant (t value N 1.96). Consequently, the unidimensionality of a measure is confirmed by statistic results in this study. Composite reliability (CR) is a common statistic test for assessing the reliability of a measure. As suggested by Bagozzi and Yi (1988), composite reliabilities should exceed 0.70. Table 7 shows that the composite reliabilities of each construct ranged between 0.69 (slightly below the recommended minimums) and 0.80, which are acceptable reliability values. The model was further assessed for convergent and discriminant validity (Table 7). Convergent validity was assessed using the following criteria: (1) individual item loading (λ) should be

significant (t-values should exceed 1.96) and exceed 0.5 [32,39,68], (2) composite reliabilities should exceed 0.70 [8,39], and (3) the average variance extracted (AVE) for each construct should exceed 0.50 [8,39]. Table 7 shows that all estimated standard loadings are statistically significant (t-values should exceed 1.96) and all items exceeded the cutoff of 0.50. Furthermore, the composite reliabilities of each construct either approached or exceeded the cutoff of 0.70 (ranged between 0.69 and 0.80), while the AVE of all constructs lay between 0.44 and 0.58. The significance of the item loading and composite reliabilities for convergent validity thus are met. Meanwhile, two of the average variance extracted scores are slightly below the recommended minimums. To assess the discriminant validity – the degree to which different constructs diverge from one another – this study followed Fornell and Larcker (1981), who recommended testing whether the AVE of each construct exceeded the correlations between two constructs [32]. Table 7 shows that the AVE of any two constructs always exceeded the correlations between two constructs. 4.2. Structural model: hypothesis and model testing The theoretical model and hypothesized relationships presented earlier were collectively tested using AMOS 5.0. Goodness of fit was tested using six common model-fit measures: chi-square/degrees of freedom, goodness-of-fit index (GFI; [40]), adjusted goodness-of-fit index, normed fit index (NFI; [11]), root mean square error of approximation (RMSEA; [19]), and comparative fit index (CFI; [10]). The results of the model fit analysis suggest that the general structural model closely fits the data (as shown in Table 8). Fig. 2 displays the standardized path coefficients, path significances, and variance explained (R2 value) by each path. The R2, the measure of percent variance explained, was 53.7% for the auction continuance intention, 38% for trust in actor, and 4.1% for shared vision. This study proposed direct links between shared vision (H1), trust in actors (H2), social interaction (H3a and H3b), and auction continuance intention. Three paths for shared vision, trust in actor, and bidding stage interaction were significant, while the path for transaction-commit stage interaction was not significant. Hypothesis 4 proposed a link between shared vision and trust in actors. The path

Table 7 Scale reliability and validity. Construct

Shared vision

Item

SV1 SV2 SV3 Trust in Trust1 actors Trust 2 Trust 3 Auction CI ACI1 ACI2 ACI3

Standardized t item loading value

Composite AVE reliability; CR

Squared correlation

0.515 0.653 0.784 0.849 0.792 0.614 0.639 0.7 0.713

0.69

0.44 0

0.80

0.58

0.23 0

0.73

0.47

0.25 0.21

SV

Pearson

Trust ACI

5.88 6.15 10.561 8.437 7.195 7.257

0 Fig. 2. SEM analysis of the research model.

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476 Table 9 Standardized direct, indirect and total effects. Shared vision Bidding stage interaction Direct effects 0.005 Indirect effects – Total effects 0.005

Trust in actor

Auction continuance intention

− 0.022 0 − 0.022

0.179 0 0.179

Transaction-commit stage interaction Direct effects 0.2 0.046 Indirect effects – 0.122 Total effects 0.2 0.168

− 0.003 0.146 0.143

Shared vision Direct effects Indirect effects Total effects

– – –

0.608 – 0.608

0.459 0.194 0.653

Trust in actor Direct effects Indirect effects Total effects

– – –

– – –

0.318 – 0.318

was positive and significant, indicating that the cognitive capital of auction community, such as actor's shared vision, increases trust in other actors. Hypotheses 5a and 5b suggested a link between high levels of social interaction and shared vision. The results show the existence of a positive and significant link between transactioncommit stage interaction and shared vision, while no such link was identified between bidding stage interaction and shared vision. Finally, hypotheses 6a and 6b suggested a link between high levels of social interaction and trust in actor. The analytical results indicated that neither transaction-commit stage interaction nor bidding stage interaction was linked to actor's trust in other actors. Table 9 lists the standardized direct, indirect and total effects of the model. The results show that transaction-commit stage interaction affected shared vision and, through it, indirectly affected trust in actor and auction continuance intention. The total standardized effects on auction continuance intention were as follows: transaction-commit stage interaction, 0.116; and shared vision, 0.653. Furthermore, the total standardized effect of transaction-commit stage interaction on trust in actor was 0.168. 5. Discussion and conclusion 5.1. Effects on continuance intention The analytical results showed that shared vision, trust and bidding stage interaction positively affect intention to continue using an auction site (Hypotheses 1, 2 and 3a), while transaction-commit stage interaction, does not affect continuance intention (Hypotheses 3b). Close examination reveals that shared vision is a stronger predictor of continuance intention than trust. Previous studies have argued that trust is the core of any business transaction or bilateral exchange [6,7,59]. However, the findings of this study indicate that the path coefficient from shared vision to continuance intention is stronger than that from trust to continuance intention (0.459 N 0.318 as shown in Fig. 2), indicating that auction community actors stress sense of community (of which shared vision is a key element) more than bilateral relationships such as trust. This study thus suggests that auction website management teams should stress the importance of fostering a shared vision in constructing a trust transaction environment, while previous studies focused on adopting a third-party trust mechanism to establish a trust and safe transaction environment. Intuitively, higher social interaction (namely, more Q&A activity or higher reputation rating) leads to higher continuance intention. However, the analytical results show that two stages of social interaction exert varied effects on continuance intention. Transac-

473

tion-commit stage interaction does not directly influence continuance intention. Rather, transaction-commit stage interaction indirectly affects continuance intention and is mediated by shared vision, while bidding stage interaction directly influences continuance intention. Comparing the above results with those of previous studies reveals some interesting patterns. First, as mentioned above, a positive comment can result from positive perceptions of repeated interactions (such as negotiations involving price, payment and logistics) and can be considered the reciprocal contact during the transaction-commit stage, while Q&A is the easiest to create and can be considered an accidental interaction or asymmetrical contact during the bidding stage. As suggested by Granovetter [37], asymmetrical contact can be treated as a weak tie while reciprocal contact can be treated as a strong tie. In light of this notion, the bidding stage interactions are analogous to weak ties and transaction-commit stage interactions to strong ties in auction communities. Strong ties tend to be more clustered and transitive, and thus encourage the fostering of community [80,81]. Our findings are consistent with this result in that transaction-commit stage interactions would enhance shared vision, which is considered critical in creating a meaningful sense of community. On the other hand, weak ties are more likely to produce new information than strong ties [23,37]. This is reflected in the fact that auction actors tend to seek new information through Q&A interactions. Auction actors with weak ties (such as a series of questions and answers) thus are more likely to have opportunities to contact various strangers, increasing their intention to engage in transactions with various strangers. However, interaction during the transaction-commit stage is not positively related with continuance intention. This phenomenon may result because interactions at this stage merely reflect how buyers and sellers rate the behavior of their counterparts in fulfilling their respective promises regarding product shipping, payment, and so on. Consequently, there does not appear to be any clear relationship with increasing future trading intentions. This phenomenon may explain why this study failed to identify any significant and direct effect of transaction-commit stage interaction on continuance intention. Second, although transaction-commit stage interaction does not directly affect continuance intention, the results of this study indicate a significant link between transaction-commit stage interaction and shared vision, which further influences continuance intention. Several studies contended that reputation system is a source of information guiding decisions related to transactions with strangers [16,28,29,51,56]. However, Ba [5] argued that reputation history (namely transaction-commit stage interaction) does not encourage future transactions due to “the lack of strong authentication of online identities”. The empirical findings of this study suggest that highly positive reputation (namely transaction-commit stage interaction) enhances the shared vision between an actor and others, and thus indirectly affect continuance intention. A conceptual direction thus can be provided regarding how to evaluate reputation system effectiveness. In other words, this result implies that simply inflating reputation ratings (or transaction number) without simultaneously elevating the other two aspects of social capital, namely shared vision and trust, may not enhance continuance intention. This implication also answers the question of why two auction sites with the same trust mechanisms may not enjoy equal success in fostering communities. 5.2. Explaining the links among three dimensions of social capital Previous studies have shown the effectiveness and associability of three dimensions of social capital — shared vision (key element of cognitive dimension), social interaction (key element of structural dimension), and trust (key element of relational dimension) [26,75,78]. However, few studies have empirically tested the relationships among these dimensions in the context of online communities,

474

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

especially in the context of transaction-oriented communities. This study tests the theoretical relationships for the three dimensions of social capital proposed by Nahapiet and Ghoshal [57]. Social interactions, including both bidding stage and transactioncommitment stage interactions, exert different effects on shared vision and trust. First, the analytical results show that bidding stage interaction affects neither shared vision nor trust in actors. As mentioned earlier, this study views Q&A-based ties as a form of social interaction during the bidding stage in auction communities. This study contends that social interaction during the bidding stage appears loosely related to enhancing sense of community in spite of it being helpful for enhancing continuance intention by obtaining new information from strangers. Additionally, examining R2 revealed that shared vision was low (0.034), suggesting that other variables may be omitted from the proposed model, or that the indices of social interaction in the proposed model may be insufficient. Future studies can examine other applicable indices of social interaction following the availability of full network structure information. Second, social interaction during the transaction-commit stage positively affects shared vision while it does not directly influence trust. Rather, shared vision mediates the influence of transactioncommit stage interaction on trust. This implies that, in the context of transaction-oriented community, feelings of trust and trustworthiness cannot be derived directly from intensive interaction (high reputation). High reputation is not necessarily a predicator of high trust. The model suggests that achieving high trust requires establishing a shared vision or sense of community to ensure alignment between individual and community values. Restated, reputation enhancement mechanisms provided by a trusted third party are ineffective in mobilizing trust unless the community embodying the mechanism has cultivated certain forms of sense of community. 5.3. Conclusion This study investigates continuance intention of online auction actors by applying social capital theory to examine social interaction. The varied dimensions of social capital embedded in auction actors influence auction continuance intention. More specifically, social interaction of bidding stage directly affects continuance intention while social interaction of transaction-commit stage does not. The results of this study also indicated that shared vision mediated the impact of transaction-commit stage interaction on trust and continuance intention. Furthermore, shared vision both directly and indirectly affected continuance intention through trust. Theoretically, the findings of this study respond to the call of Lamb for ICT research to view users as social actors in various social contexts [46]. As noted above, auction continuance intention is driven by the interrelation among the three dimensions of social capital, which enhances auction continuance intention. While there are numerous studies addressing how to conceptualize these three dimensions of social capital, there is little empirical evidence regarding how the interplay among the three dimensions of social capital influences continuance intention for virtual communities. The findings of this study provide valuable insights regarding how social capital can be fostered to achieve management objectives in an online community such as an online auction site. Some implications of this study are discussed below. First, previous studies rarely specified the relationships among the three dimensions of social capital. The empirical evidence presented in this study showed that elements in the three dimensions of social capital had different effects in terms of mobilizing social capital. Trust directly influences auction continuance intention and mediates the influence of shared vision on auction continuance intention; meanwhile, shared vision mediated the influence of social interaction on continuance intention and trust. Social interaction, arising from either Q&A or reputation rating, facilitated more referral, access, and

exposure within the community. The interaction during the transaction-commit stage (that is, strong social interaction) may lead to a shared vision and language, or to an enhanced sense of community, while the interaction during the bidding stage (that is, weak strong ties) may provide auction actors increased future interaction opportunities. That is, strong social interaction can be considered an antecedent indicator of shared vision and trust, a conclusion consistent with the claim of Bourdieu [17] that interaction is a precondition for developing and maintaining dense social capital. These findings also suggest that management of auction websites should be aware of the different impacts of various stages of social interaction and adopt appropriate mechanisms to encourage social interaction. Second, the research model proposed in this study indicated that interaction within a social context could trigger social capital and influence continuance intention. The influence process was initiated by individual interaction ties (e.g. positive reputation and Q&A interaction) embedded within social contexts. Some interaction ties helped establish dyadic relationships for cultivating shared vision and trust. Consequently, social capital within virtual communities is created with the development of shared language, shared vision, and sense of community. The effect of social capital thus results in influence continuance intention for online auction websites. In practice, since positive transaction record or reputation ratings do not automatically lead to high continuance intention, website designs and institutional mechanisms alone may be insufficient to foster online auction community growth. Auction website management thus should place greater stress on improving the management of these mechanisms to build a social context that enhances social capital. For example, two equally reputable auction actors from two different auction websites adopting the same reputation mechanism may not share the same level of intention to continue interacting. Furthermore, the interaction experiences of each actor may result in different levels of trust or shared vision, potentially reducing or increasing actor continuance intention. Essentially, reputation level alone is not an effective indicator of continuance intention, and hinges on the level of social capital mobilized. In conclusion, reputation should be combined with other indexes of social interaction to yield a superior predictor of actor behavior. Furthermore, web site managers must be ever cognizant of interactions among auction actors and the effects of social capital on auction continuance intention. 5.4. Limitations and future research directions This study suffers several limitations. First, it is difficult to measure social interaction owing to the lack of full network information and the inability to investigate the process of social capital development or the manner in which the network structure changes over time. This study thus relied on the exposed connectivity-based interaction ties as indicators of social interaction, but ignored interaction content. Future studies thus should consider the dynamic nature of network structuring, using more longitudinal data and additional measures of interaction ties as indicators of social interaction. Alternatively, future studies could also benefit from examining different dependent variables based on various communication activities, such as on-line discussion forums. Second, this study assumed that social interaction/ties and social context were constructed in an on-line context. However, offline social evolvement may influence social actors. This study did not investigate individuals whose social ties are constructed primarily by offline interaction and for whom online interactions are secondary. Future studies should examine whether on-line and off-line social context and social capital affect continuance intention differently. A related question is whether the social capital activation model applies to different non-business transaction practices, such as those focused on hobbies or open source development teams.

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

Finally, the sampling method is another limitation of this study. Owing to the nature of online auctions (such as the use of pseudonyms and a lack of demographic information), it is hard to utilize standard sampling and estimation techniques. This study thus contends that auction communities are a “hard-to-reach” or “hidden” population and employs a snowball sampling method. This approach inevitably encounters problems, such as missing individuals who are isolated from social networks, and oversamples the more privileged segments of the population. Owing to these biases, it is recommended that future studies should utilize multiple sampling to check the validity of the results, or alternatively use larger sample sizes to reduce bias.

References [1] M. Acevedo, Network capital: an expression of social capital in the network society, Proceedings of the UOC Virtual Conferences Fall 2003, 2003. [2] M.K. Ahuja, D.F. Galletta, K.M. Carley, Individual centrality and performance in virtual R&D groups: an empirical study, Management Science 49 (1) (2003) 21–38. [3] R. Algesheimer, U.M. Dholakia, A. Herrmann, The social influence of brand community: evidence from European car clubs, Journal of Marketing 69 (3) (2005) 19–34. [4] K. Anirudh, Active Social Capital: Tracing The Roots of Development and Democracy, Columbia University Press, 2002. [5] S. Ba, Establishing online trust through a community responsibility system, Decision Support Systems 31 (3) (2001) 323–336. [6] S. Ba, A.P. Paul, Evidence of the effect of trust building technology in electronic markets: price premium and buyer behavior, MIS Quarterly 26 (3) (2002) 243–268. [7] S. Ba, A.B. Whinston, H. Zhang, Building trust in online auction markets through an economic incentive mechanism, Decision Support Systems 35 (3) (2003) 273–286. [8] R.P. Bagozzi, Y. Yi, On the evaluation of structural equation models, Journal of Academy of Marketing Science 16 (1) (1988) 74–94. [9] J.E. Bailey, S.W. Pearson, Development of a tool for measuring and analyzing computer user satisfaction, Management Science 29 (5) (1983) 519–529. [10] P.M. Bentler, Comparative fit indexes in structural models, Psychological Bulletin 107 (2) (1990) 238–246. [11] P.M. Bentler, D.G. Bonett, Significant tests and goodness of fit in the analysis of covariance structures, Psychological Bulletin 88 (3) (1980) 588–606. [12] S. Beugelsdijk, T.v. Schaik, Social capital and growth in European regions: an empirical test, European Journal of Political Economy 21 (2) (2005) 301–324. [13] A. Bhattacherjee, An empirical analysis of the antecedents of electronic commerce service continuance, Decision Support Systems 32 (2) (2001) 201–214. [14] A. Bhattacherjee, Understanding information systems continuance: an expectation–confirmation model, MIS Quarterly 25 (3) (2001) 351–370. [15] A. Blanchard, T. Horan, Virtual communities and social capital, Social Science Computer Review 16 (3) (1998) 293–307. [16] G.E. Bolton, E. Katok, A. Ockenfels, How effective are electronic reputation mechanisms? An experimental investigation, Management Science 50 (11) (2004) 1587–1602. [17] P. Bourdieu, The forms of capital, in: J.C. Richardson (Ed.), Handbook of Theory and Research for the Sociology of Education, New York, Greenwood, 1986. [18] J. Boyd, In community we trust: online security communication at eBay, Journal of Computer-Mediated Communication 7 (3) (2002) Electronic Edition. [19] M.W. Browne, R. Cudeck, Single sample cross-validation indices for covariance structures, Multivariate Behavioral Research 24 (4) (1989) 445–455. [20] C.-A. Bryant, D. Norris, Measurement of social capital: the Canadian experience, Proceedings of the UK ONS International Conference on Social Capital Measurement in London, 2002. [21] E. Brynjolfsson, M. Smith, Frictionless commerce? A comparison of Internet and conventional retailers, Management Science 46 (4) (2000) 563–585. [22] S.M. Burroughs, L.T. Eby, Psychological sense of community at work: a measurement system and explanatory framework, Journal of Community Psychology 26 (6) (1998) 509–532. [23] R.S. Burt, Structural Holes: The Social Structure of Competition, Harvard University Press, MA, 1992. [24] G. Cai, P.R. Wurman, Monte Carlo approximation in incomplete information, sequential auction games, Decision Support Systems 39 (2) (2005) 153–168. [25] C.-M. Chiu, M.-H. Hsu, S.-Y. Sun, T.-C. Lin, P.-C. Sun, Usability, quality, value and elearning continuance decisions, Computers & Education 45 (4) (2005) 399–416. [26] C.-M. Chiu, M.-H. Hsu, E.T.G. Wang, Understanding knowledge sharing in virtual communities: an integration of social capital and social cognitive theories, Decision Support Systems 42 (3) (2006) 1872–1888. [27] F.D. Davis, R.P. Bagozzi, P.R. Warshaw, User acceptance of computer technology: a comparison of two theoretical models, Management Science 35 (8) (1989) 982–1003. [28] C. Dellarocas, Reputation mechanism design in online trading environments with pure moral hazard, Information Systems Research 16 (2) (2005) 209–230. [29] C. Dellarocas, How often should reputation mechanisms update a trader's reputation profile? Information Systems Research 17 (3) (2006) 271–285.

475

[30] W.H. DeLone, E.R. McLean, Information systems success: the quest for the dependent variable, Information Systems Research 3 (1) (1992) 60–95. [31] G. Fischer, E. Scharff, Y. Ye, Fostering social creativity by increasing social capital, in: M. Huysman, V. Wulf (Eds.), Social Capital and Information Technology, MIT Press, Cambridge, MA, 2004. [32] C. Fornell, D.F. Larcker, Evaluating structural equation models with unobservables and measurement error, Journal of Marketing Research 18 (1981) 39–50. [33] L. Freese, Social interaction: what is it? Proceedings of the 82nd Annual Meeting of the American Sociological Association, Chicago, IL, 1987. [34] V.J. Garc'ia-Morales, F.J. Llorens-Montes, A.J. Verdu'-Jover, Antecedents and consequences of organizational innovation and organizational learning in entrepreneurship, Industrial Management & Data Systems 106 (1) (2006) 21–42. [35] D. Gefen, D.W. Straub, Managing user trust in B2C e-services, e-Service Journal 2 (2) (2003) 7–24. [36] M. Granovetter, Economic action and social structure: the problem of embeddedness, American Journal of Sociology 91 (3) (1985) 481–510. [37] M. Granovetter, Getting a Job: A Study of Contacts and Careers, University Of Chicago Press, Chicago, 1995. [38] R. Gulati, Does familiarity breed trust? The implications of repeated ties for contractual choice in alliances, Academy of Management Journal 38 (1) (1995) 85–112. [39] J.F. Hair, R.L. Tatham, R.E. Anderson, W. Black, Multivariate Data Analysis, 5th ed. Prentice-Hall, Englewood Cliffs, NJ, 1998. [40] K. Jöreskog, D. Sörbom, LISREL 8: Structural Equation Modeling with the SIMPLIS Command Language, Erlbaum, Hillsdale, NJ, 1993. [41] A. Jøsang, R. Ismail, C. Boyd, A survey of trust and reputation systems for online service provision, Decision Support Systems 43 (2) (2007) 618–644. [42] S. Jones, Community-based ecotourism: the significance of social capital, Annals of Tourism Research 32 (2) (2005) 303–324. [43] D.J. Kim, D.L. Ferrin, H.R. Rao, A study of the effect of consumer trust on consumer expectations and satisfaction: the Korean experience, Proceedings of the Proceedings of the 5th International Conference on Electronic Commerce, Pittsburgh, Pennsylvania, 2003, pp. 310–315. [44] J. Koh, Y.-G. Kim, Sense of virtual community: a conceptual framework and empirical validation, International Journal of Electronic Commerce 8 (2) (2003) 75–93. [45] R. Konrad, EBay launches ‘neighborhoods’ feature, http://www.usatoday.com/ tech/products/2007-10-09-ebay-neighborhood_N.htm, as of 2007, (2007). [46] R. Lamb, R. Kling, Reconceptualizing users as social actors in information systems research, MIS Quarterly 27 (2) (2003) 197–235. [47] R. Landry, N. Amara, M. Lamari, Social capital, innovation and public policy, ISUMA, Canadian Journal of Policy Research 2 (1) (2001) 63–72. [48] E.L. Lesser, J. Storck, Communities of practice and organizational performance, IBM Systems Journal 40 (4) (2001). [49] J. Liao, H. Welsch, Roles of social capital in venture creation: key dimensions and research implications, Journal of Small Business Management 43 (4) (2005) 345–362. [50] N. Lin, Building a network theory of social capital, Connections 22 (1) (1999) 28–51. [51] Z. Lin, D. Li, B. Janamanchi, W. Huang, Reputation distribution and consumer-toconsumer online auction market structure: an exploratory study, Decision Support Systems 41 (2) (2006) 435–448. [52] J. Lynch, P. Due, C. Muntaner, G.D. Smith, Social capital — is it a good investment strategy for public health? Journal Epidemiology Community Health 54 (2000) 404–408. [53] G. Marwell, P. Oliver, The Critical Mass in Collective Action, Cambridge University Press, New York, 1993. [54] V. McKinney, K. Yoon, F.M. Zahedi, The measurement of web-customer satisfaction, An Expectation and Disconfirmation Approach 13 (3) (2002) 296–315. [55] D.W. McMillan, D.M. Chavis, Sense of community: a definition and theory, Journal of Community Psychology 14 (1) (1986) 6–23. [56] M.I. Melnik, J. Alm, Does a seller's ecommerce reputation matter? Evidence from eBay auctions, Journal of Industrial Economics 50 (3) (2002) 337–349. [57] J. Nahapiet, S. Ghoshal, Social capital, intellectual capital and the organizational advantage, Academy of Management Review 23 (2) (1998) 242–266. [58] M. Paldam, G.T. Svendsen, An essay on social capital: looking for fire behind the smoke, European Journal of Political Economy 16 (2) (2000) 339–366. [59] P.A. Pavlou, D. Gefen, Building effective online marketplaces with institutionbased trust, Information Systems Research 15 (1) (2004) 37–59. [60] E.J. Pinker, A. Seidmann, Y. Vakrat, Managing online auctions: current business and research issues, Management Science 49 (11) (2003) 1457–1484. [61] C.E. Pollack, O. Knesebeck, Social capital and health among the aged: comparisons between the United States and Germany, Health & Place 10 (4) (2004) 383–391. [62] J. Preece, Supporting community and building social capital, Communications of the ACM 45 (4) (2002) 37–39. [63] R.D. Putnam, R. Leonard, R.Y. Nanetti, Making Democracy Work: Civic Traditions in Modern Italy, Princeton University Press, Princeton, NJ, 1993. [64] F.F. Reichheld, P. Schefter, E-loyalty: your secret weapon on the web, Harvard Business Review 78 (4) (2000) 105–113. [65] D.M. Rousseau, S.B. Bitkin, R.S. Burt, C. Camerer, Not so different after all: a crossdiscipline view of trust, Academy of management Review 23 (3) (1998) 393–404. [66] M.A. Royal, R.J. Rossi, Individual-level correlates of sense of community: findings from workplace and school, Journal of Community Psychology 24 (4) (1996) 395–416. [67] M.J. Salganik, D.D. Heckathorn, Sampling and estimation in hidden populations using respondent-driven sampling, Sociological Methodology 34 (2004) 193–239. [68] D. Samson, M. Terziovski, The relationship between total quality management practices and operational performance, Journal of Operations Management 17 (4) (1999) 393–409.

476

J.-C. Wang, M.-J. Chiang / Decision Support Systems 47 (2009) 466–476

[69] J.K. Scott, T.G. Johnson, Bowling alone but online together: social capital in ecommunities, Journal of the Community Development Society 36 (1) (2005) 1–18. [70] D.L. Shapiro, B.H. Sheppard, L. Cheraskin, Business on a handshake, Negotiation Journal 8 (4) (1992) 365–377. [71] S.S. Smith, Don't put my name on it: social capital activation and job-finding assistance among the Black urban poor, American Journal of Sociology 111 (2005) 1–57. [72] K. Stanoevska-Slabeva, Toward a community-oriented design of Internet platforms, International Journal of Electronic Commerce 6 (3) (2002) 71–95. [73] T.J. Strader, S.N. Ramaswami, The value of seller trustworthiness in C2C online markets, Communication of the ACM 45 (12) (2002) 45–49. [74] A. Tiwana, A.A. Bush, Continuance in expertise-sharing networks: a social perspective, IEEE Transactions on Engineering Management 52 (1) (2005). [75] W. Tsai, S. Ghoshal, Social capital and value creation: the role of intrafirm networks, Academy of Management Journal 41 (4) (1998) 464–476. [76] V. Venkatesh, F.D. Davis, A theoretical extension of the technology acceptance model: four longitudinal field studies, Management Science 46 (2) (2000) 186–204. [77] S. Walczak, D.G. Gregg, J. Berrinberg, Market decision making for online auction sellers: profit maximization versus socialization perspective, Journal of Electronic Commerce Research 7 (4) (2006) 199–220. [78] M.M. Wasko, S. Faraj, Why should i share? Examining social capital and knowledge contribution in electronic networks of practice, MIS Quarterly 29 (1) (2005) 35–57. [79] B. Wellman, Networks in the Global Village: Life in Contemporary Communities, Westview Press, 1999. [80] D.R. White, M. Houseman, The navigability of strong ties: small worlds, tie strength and network topology, Complexity 8 (1) (2002) 72–81.

[81] K.P. Wilkinson, The Community in Rural America, Greenwood Press, New York, 1991. [82] J.R. Wolf, W. Muhanna, Adverse selection and reputation systems in online auctions: evidence from eBay motors, Proceedings of the Twenty-Sixth International Conference on Information Systems, 2005, pp. 847–858. [83] L.G. Zarker, Production of trust: institutional source of electronic structure, 1984– 1920, Research in Organizational Behavior 8 (1986) 55–111. Jyun-Cheng Wang is an Associate Professor of MIS in Institute of Service Science at the National Tsing-Hua University, Taiwan, ROC. He holds a Ph.D. in Management of Information Systems from the University of Wisconsin-Madison. His research interests are focused on applying social network analysis for technology-mediated interactions. He has published papers on such topics as online auctions, internet marketing, and the patterns of development of virtual communities. Ming-Jiin Chiang is a Ph.D. candidate in the Department of Information Management at the National Chung Cheng University, Taiwan, ROC. He received M.S. degree in Information Management from National Chung Cheng University in 1997, and works as a faculty member in Department of Information Management, TaTung Institute of Commerce and Technology, Chiayi City, Taiwan (R.O.C.). His doctoral research is focused on social interaction in the virtual community, social network analysis, research network. His research has appeared in JAIS Sponsored Theory Development Workshop Following ICIS 2006.