A multilevel model for effects of social capital and knowledge sharing in knowledge-intensive work teams

International Journal of Information Management 33 (2013) 780–790

Contents lists available at ScienceDirect

International Journal of Information Management journal homepage: www.elsevier.com/locate/ijinfomgt

A multilevel model for effects of social capital and knowledge sharing in knowledge-intensive work teams Yan Yu a,1 , Jin-Xing Hao b,∗ , Xiao-Ying Dong c,2 , Mohamed Khalifa d a

Information Systems Department, Information School, Renmin University of China, 59 Zhongguancun Street, Haidian District, Beijing 100872, PR China School of Economics and Management, BeiHang University, Beijing 100191, PR China Guanghua School of Management, Peking University, 5 Yiheyuan Road, Haidian District, Beijing 100871, PR China d Al Ghurair Univeristy, United Arab Emirates b c

a r t i c l e

i n f o

Article history: Available online 6 July 2013 Keywords: Knowledge management Knowledge sharing Social capital Social network Multilevel approach

a b s t r a c t Although it is a widely held belief that social capital facilitates knowledge sharing among individuals, there is little research that has deeply investigated the impacts of social capital at different levels on an individual’s knowledge sharing behavior. To address this research gap, this study combines a multilevel approach and an optimal network configuration view to investigate the multilevel effects of social capital on individuals’ knowledge sharing in knowledge intensive work teams. This study makes a distinction between the social capital at the team-level and that of social capital at the individual level to examine their cross-level and direct effects on an individual’s sharing of explicit and tacit knowledge. A survey involving 343 participants in 47 knowledge-intensive teams was conducted for testing the multilevel model. The results reveal that social capital at both levels jointly influences an individual’s explicit and tacit knowledge sharing. Further, when individuals possess a moderate betweenness centrality and the whole team holds a moderate network density, team members’ knowledge sharing can be maximized. These findings offer a more comprehensive and precise understanding of the multilevel impacts of social capital on team members’ knowledge sharing behavior, thus contributing to the social capital theory, as well as knowledge management research and practices. © 2013 Elsevier Ltd. All rights reserved.

1. Introduction Knowledge is a crucial asset for organizations to gain sustainable competitive edge (Grant, 1996). Organizational competitiveness is rooted in the mobility of knowledge that is realized through knowledge sharing and transfer. It has been shown that knowledge sharing provides individuals, work teams and organizations with the opportunity to improve the work performance as well as create new ideas and innovations (Cumming, 2004). To enhance knowledge sharing, modern organizations have attempted to adopt decentralized, flat structures such as work team settings. One merit of team-based organizational design lies in its flexible social structure (Argote, McEvily, & Reagans, 2003), in which individuals not only benefit from the close relationships within teams but also have

∗ Corresponding author at: School of Economics and Management, Beihang University, 37 Xueyuan Road, Haidian District, Beijing 100191, China. Tel.: +86 10 82339735. E-mail addresses: [email protected] (Y. Yu), [email protected], [email protected] (J.-X. Hao), [email protected] (X.-Y. Dong), [email protected] (M. Khalifa). 1 Tel.: +86 10 8250 0906. 2 Tel.: +86 10 6275 7290. 0268-4012/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijinfomgt.2013.05.005

the flexibility to establish relationships with external sources for acquiring and providing knowledge. However, little research has deeply investigated the specific network structure of a work team that fosters the internal and external social capital and largely influences team members’ knowledge sharing behavior. Prior social capital and social network research has largely limited the analytic unit to single levels, thus neglecting the cross-level effect, i.e. the effect of the aggregated overall network on individual behavior. Most prior studies provide evidence of the impacts of an individual’s personal relationships on their perceived organizational support (Zagenczyk, Scott, Gibney, Murrell, & Thatcher, 2010), boundary-spanning (Xiao & Tsui, 2007), mobility (Podolny & Baron, 1997), knowledge sharing intention (Hau, Kim, Lee, & Kim, 2013) and behavior (Yang & Farn, 2009). Although an individual’s social networking is important, disregarding the contextual influence at a higher level may lead to an under-estimation of the impacts of the social capital and network as a whole, while simultaneously overemphasizing the effects of discrete personal networks. While some recent research addresses the cross-level effect of social capital on knowledge transfer or sharing, their analytic unit is focused on dyad rather than the team network that would expand as the interpersonal relationships develop. For example, Kang and Kim (2010) take a social network perspective to examine

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

the direct and moderating effects of the dyadic and individual level antecedents of knowledge transfer in dyads. Wei, Zheng, and Zhang (2011) also find that the team network structure may intervene on an individual’s network impact on knowledge transfer in dyads. Moreover, Phelps, Heidl, and Wadhwa (2012) conduct a research review of articles on knowledge networks and find a consistency in terms of how egos interpersonal networks increase their knowledge transfer and absorption, as well as the conflict resulting from the effects of ego networks and the whole network structures on knowledge outcome. Their comprehensive review suggests that the trade-off between ego networks and the whole network should be further examined. Thus, it is the worthwhile aim of this study to explore the direct and cross-level effects of social capital, which stem from social networks at different levels on individual knowledge sharing from the multilevel approach and the optimal network configuration perspective. As a collection of resources, social capital resides in both individuals and the whole collective that individuals form into (Nahapiet & Ghoshal, 1998). Social capital can be possessed by the units in different levels. An individual’s social capital is shaped upon his/her advantageous positions in the situated network, wherein he/she has mutual understanding and trusting relationships with his/her colleagues. The overall network, such as a work team, an organization, or a society with a broader boundary, also possesses social capital. Hence, it is necessary to conceptually distinguish social capital at a higher level (collective level) from that of social capital at the individual level. This is particularly important in proving the cross-level influences of the collective capital on an individual’s behavior, such as knowledge sharing. The collective social capital is over and beyond individual social capital. According to the nested structuration theory, teams provide the most immediate contextual environment for individuals and thus directly influence the behavior of the nested members (Hoegl, Parboteeah, & Munson, 2003; Perlow, Gittell, & Katz, 2004). Accordingly, this study constrained the upper side of social capital to the team level and investigated its effects on individual sharing behavior. To account for the effects of both team and individual levels of social capital on team members’ knowledge sharing behavior, a multilevel approach was preferred. This approach helps to bridge the gap in previous single-level research and offers a new perspective to reconcile the aforementioned debates in social network literature, contributing to the social capital and social network theories. The research conceptually distinguishes social capital into two levels (team-level and individual-level social capital) and further empirically illustrates the distinct roles of social capital at different levels that play in team members’ explicit and tacit knowledge sharing. Further, the research aims to verify the appropriateness of configuring optimal social networks for both teams and individuals because there is possibility that the relationships between social networks and knowledge sharing are curvilinear. This study combines a multilevel view and an optimal view of social networking, offering a more comprehensive and precise understanding of the multilevel impacts of social capital on individual knowledge sharing behavior. The impact of multilevel social capital is expected to inform management that managerial actions formulated from multiple levels of perspective is beneficial for the enhancement of team members’ knowledge sharing and other cooperative behaviors. The provision of managerial actions at both levels will yield significant managerial impacts on individual behavior. The remainder of the paper is organized as follows: Section 2 provides the theoretical foundation and research model; Section 3 describes the details of the research methodology; Section 4 illustrates and interprets the results of data analysis; and

781

finally, Section 5 concludes the paper with limitations and implications.

2. Theoretical foundation and research model 2.1. Social capital and social network theories Social capital is “the sum of the actual and potential resources embedded within, available through, and derived from the network of relationships possessed by an individual or social unit” (Nahapiet & Ghoshal, 1998, p. 243). Social capital is comprised of both the structure of the network and the potential resources that may be mobilized through the network. Accordingly, social capital, as a set of resources rooted in networking relationships, can be decomposed into three distinct facets: structural capital, cognitive capital, and relationship capital. Structure capital describes the impersonal configuration of linkages among a social group of people; cognitive capital is derived from the shared representations, interpretations, and meaning among the members who are located in the social group; and relational capital refers to the affective nature of the networking relationships where the situated members have a strong identification toward this particular social group, perceive an obligation of participation, and abide by cooperative norms (Putnam, 1993). The structural capital, together with the embedded cognitive and relational capital, supplies motives for individuals to act collectively and share knowledge. Considering the nested nature of work teams where individuals situate, this study uses a multilevel approach to examine the influences of social capital residing in both individual members and the whole team. An individual’s social interactions help to establish the team-level social capital, which in turn influences individual behavior. The social capital of a team is derived from a broader network which is defined by the social boundary of the team instead of by its formal boundary. This broader network not only includes the team members’ internal network ties, but also their external ties within the organization. In the bottom-up process of social capital formation, individuals who possess certain positions in the broad team network, construct shared cognition with other members, seek to transmit the team identity into an individual identity, and establish an emotional attachment to the particular work team. Consistent with prior research (e.g. Reagans & McEvily, 2003; Wasko & Faraj, 2005), this study postulates that individually held social capital, which is shaped by the individual’s position centrality, perceived shared cognition, and affective commitment, supplies crucial stimuli for him/her to engage in knowledge sharing within the work teams. An individual’s position centrality reflects his/her structural capital. The individual who is centrally embedded in a collective has a relatively high proportion of direct ties to other members, and thus has advantages in accessing and contributing knowledge (Wasko & Faraj, 2005). An individual’s cognitive capital develops as he/she interacts with individuals sharing the same practice during a given period of time, learns the skills, knowledge, specialized discourse, and norms of the practice, and finally builds shared cognition within teams (Nahapiet & Ghoshal, 1998). As Coleman (1990) suggests, the main function of the relational capital is to facilitate an individual’s actions within the structure. An individual with affective commitment to the collective will convey a sense of responsibility to help others on the basis of shared membership. At a higher level, team social capital as an aggregate of the compositions (team members) can exert important impacts on individual behavior (Hansen, Mors, & Lovas, 2005; Oh, Chung, & Labianca, 2004; Oh, Labianca, & Chung, 2006). This study views this as a top-down influential process. Team social capital is reflected by the overall connectivity of the broader network of team

782

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

Team-level social capital Network density Cognition commonality

H2

Cooperative norms

H4 H6

Individual-level social capital Betweenness centrality Shared cognition

H1 H3

Knowledge sharing behavior Explicit knowledge Tacit knowledge

H5 Affective commitment Fig. 1. Research model.

members’ ties (i.e. network density), the cognition commonality and the emergence of cooperative norms within the teams. According to the nested structuration theory (Perlow et al., 2004), the circumstance in a team emerges from individuals’ interaction and directly influences the individuals’ actions in turn. Thus, the research proposes a research model to investigate the impacts of individual-level and team-level social capital on knowledge sharing in Fig. 1. Further elaborations on the model are detailed in the following subsections. 2.2. Structural capital and knowledge sharing The positions of people in a network allow them different scales of opportunities and abilities to provide knowledge to, as well as receive knowledge from, others. In social network literature, two main network configurations are suggested: closure relationships (Coleman, 1990) and bridging relationships with structural holes (Burt, 1992). In the closure network mechanism, individuals connected by strong ties benefit from the embeddedness of the relationships in their closed social group, whereas in the bridging network mechanism, bridging ties that connect otherwise disconnected individuals enjoy information and control benefits. The gaps between disconnected individuals are referred to as structural holes (Burt, 1992). This study posits that the fundamental difference between the two mechanisms of social networks is rooted in their focus at different levels. As Adler and Kwon (2002) observe, the closure network mechanism emphasizes that the overall connections among individuals in a collective give the collective cohesiveness, thereby facilitating the pursuit of collective goals; in contrast, the bridging network mechanism highlights a focal actor’s advantageous position in a collective that leads to individual benefits. Reagans and McEvily (2008) also postulate that the opposite logic of the two network mechanisms could be caused by the fact that network structures are operationalized at different levels. In order to distinguish the structural capital at different levels, this study argues that the individual-level structural capital is shaped by an individual’s betweenness centrality, whereas the team-level structural capital is reflected by the density of the broad team network. Individual structural capital and knowledge sharing. An individual’s structural capital is shaped by his/her position in a particular network. Network centrality describes the degree in which the actor is important in the network. Several types of network centralities are widely noted, such as degree centrality and betweenness centrality. Among them betweenness centrality indicates how powerful an actor is based on the number of otherwise disconnected actors this actor is connected to (Wasserman & Faust, 1994).

Removal of this focal actor results in communication issues for other network actors (Zhu & Watts, 2010), thus this study uses betweenness centrality as a critical measure for individual structural capital, as done in several prior studies (Mehra, Kilduff, & Brass, 2001; Zhu & Watts, 2010). From the bridging view of social network, people with exclusive relation to otherwise disconnected people tend to gain greater benefits (Mehra et al., 2001). Team members can establish bridges among the unconnected members within the team and bridges between the internal members and external members in the organization. Oh et al. (2006) categorize the bridging ties into intra-team bridging ties and inter-team bridging ties, respectively. Central individuals with a high proportion of ties to other members have more relationships to draw upon when obtaining knowledge and resources. It is also easier for central people to deliver knowledge and resources to others. They act as gatekeepers for information that flows through the network. Intra-team bridging members who are aware of the structural holes within the team are more likely to recognize the need for discussion and therefore are more likely to share knowledge with the team members to address knowledge gaps. Inter-team bridging members in a broader network range are more likely to import non-redundant knowledge to the internal members. Additionally, Reagans and McEvily (2003) posit that the individuals’ engagement in boundary spanning improves their ability to convey complex ideas across distinct bodies of knowledge, thus bridging members may have a higher level of self-efficacy of knowledge sharing. Intuitively, this study stipulates that individual members with a certain level of betweenness centrality in an emerging team network have a greater capacity to transmit knowledge to one another, resulting in a smooth knowledge flow within the team. Further, the study argues for an inverted U-shaped relationship of individual betweenness centrality and their knowledge sharing, different from Oh et al.’s (2006) proposition that bridging ties always bring positive social capital resources. Not all networking is desirable, since development and maintenance of relationships can be time-consuming and may undermine team cohesion (Alderfer, 1976; Hoegl et al., 2003). Extremely high betweenness centrality of an individual mainly arises from his/her enthusiasm in external boundary spanning and results in many sparse sub-networks. Boundary spanners might be particularly susceptible to role conflict that arises from differing and inconsistent expectations among multiple constituencies (Podolny & Baron, 1997). Such a diverse and disconnected network exposes the highly betweenness-centered members to conflicting preferences and allegiances (Coleman, 1990), resulting in these members not only being less able to develop a coherent team identity, but also show less intention to contribute knowledge in the team. Thus, a network, in which a large portion of the members have a moderate level of betweenness centrality, facilitates knowledge sharing. The aforementioned justifies the following hypothesis: Hypothesis 1. The relationship between an individual’s structural capital in a team (betweenness centrality) and his/her knowledge sharing in the team is in an inverted U-shape. Team structural capital and knowledge sharing. Team structural capital is shaped by the degree of closeness of the particular network. Network density describes the overall connectivity in a social network, reflecting a team structure. In a closure social network with high density, bounded solidarity, strong trust, reciprocity, and sanctions against self-serving behaviors are expected (Coleman, 1990). Social cohesion can exert a positive effect on knowledge sharing, primarily through influencing the willingness of individuals to devote time and effort to assist, as well as learn from others (Reagans & McEvily, 2003). Nurturing strong reciprocity could be one remedy for social loafing; and the bounded solidarity of

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

closure relationships and the conformity to agree upon norms could result in more efficient and effective self sanctioning and the reduction of opportunistic behavior. Hence, a closure team network, in which individuals are trusting and face less uncertainty, results in an increased willingness among individual members to share knowledge within the team. Despite the positive consequences of the closure networking mechanism, this study does not suggest a simple positive linear relationship between the team closure and individual members’ knowledge sharing behavior. On the contrary, the excessive density of a team may constrain its extroversion, and may therefore exert a negative effect on the individuals’ knowledge sharing. Strong closure teams can constrain individual members’ contacts with external individuals, thus restricting access to more varied resources and innovative information that cannot be found within the closed team (Hansen et al., 2005; Oh et al., 2006). In this sense, excessive density of the team network may reduce the possibility for inter-team bridging relationships and valuable knowledge inflow into the team. Intra-team knowledge sharing may decrease, owing to the redundancy of knowledge. Also, Oh et al. (2004, 2006) point out that positive in-group bias and negative out-group bias, which possibly result from strong closure relationships in a team, limit the absorption of innovative external knowledge beyond the team. While an excessively dense team network can exert a negative influence on individual members’ valuable knowledge sharing with other members, the team network with a moderate level of density can maximize the valuable intra-team knowledge sharing. Thus, this study argues for an inverted U-shaped relationship between the team network density and team members’ sharing of knowledge. Accordingly, the following hypothesis is presented: Hypothesis 2. The relationship between a team’s structural capital (network density) and the nested individual’s knowledge sharing in the team is in an inverted U-shape.

2.3. Cognitive capital and knowledge sharing Individual shared cognition and knowledge sharing. Individual shared cognition refers to an individual member’s perception of the degree in which his/her team members share similar cognitive structure. The cognitive structure includes the task- and teamrelated knowledge, values, philosophies, and problem-solving approaches. The perception of interpersonal similarity stimulates the individual’s homophily behavior, i.e. the tendency to interact with similar others (Makela, Kalla, & Piekkari, 2007). According to the self-categorization theory, the perceived similarity arouses the cognitive categorization and creates opportunities for attraction from one another (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). People are more willing to share knowledge with those who hold a similar attitude, philosophy, and experience and tend to agree with them (Darr & Kurtzberg, 2000). Previous studies show that the similarity-based connections that nourish team members’ active interaction smoothen the knowledge flow within the team, because the perceived similarity reduces an individual’s psychological discomfort and conflict arising from their cognitive or emotional disparity (Borgatti & Cross, 2003; Brass, Galaskiewicz, Greve, & Tsai, 2004; Makela et al., 2007). In contrast, people will experience internal conflict and cognitive dissonance when they are faced with information that is not consistent with their own reality (Nelson & Cooprider, 1996). The lack of perceived similarity will impede individuals from engaging in knowledge sharing with their counterparts. Thus, individual members are more likely to share their knowledge with one another when they believe they are similar or cognitive equivalent (Zagenczyk et al., 2010). Accordingly, this study proposes the hypothesis below:

783

Hypothesis 3. An individual’s perceived shared cognition with other members in a team will enhance his/her knowledge sharing in the team. Team cognition commonality and knowledge sharing. As opposed to the individual’s perceived shared cognition, team-shared cognition refers to the factual overlap or commonality among all team members’ cognitive structures, which will hereafter be referred to as team cognition commonality in this study. The team cognition commonality is an aggregation of individuals’ shared cognitions to the team level, indicated by the agreement of the perceived shared cognition in a group in terms of the task- and team-related knowledge, values, philosophy, methodology, and so forth. The cognitive commonality at the team level reduces individuals’ cognitive load and calculative effort when they intend to share knowledge with others. Meaningful knowledge sharing requires at least some level of shared understanding (e.g. shared language, methodology) and mutual awareness of dealing with tasks (Nahapiet & Ghoshal, 1998). The cognition commonality helps individual members to predict the needs of tasks and the needs of the team, to anticipate the expectation and behavior of others, to adapt to changing demands, and to coordinate activities with one another successfully (Klimoski & Mohammed, 1994; Mohammed & Dumville, 2003). When a team lacks of shared knowledge base, expectations, and realities of individual members, the common team goals may also disappear; therefore, individual members may become more distant from one another and less willing to share knowledge. Thus, the cognition commonality at a team level produces an additive effect on promoting individuals’ knowledge sharing engagement in the team, which is beyond the effect of individuals’ own perceptions of the similarity. Thus, this study hypothesizes that: Hypothesis 4. A higher level of cognition commonality within a team will increase the nested individuals’ knowledge sharing in the team. 2.4. Relational capital and knowledge sharing Individual affective commitment and knowledge sharing. Affective commitment is defined as the emotional significance that individual members attach to the membership in their work teams (Van der Vegt & Bunderson, 2005), and is a result of identification. Team identification is the merger of the self and the team, with people defining themselves in terms of their group membership. Social identification nurtures one’s motivation to share knowledge; in contrast, distinct and contradictory identities within communities set up barriers to knowledge sharing (Nahapiet & Ghoshal, 1998). The affective commitment, as an emotional involvement with a particular team, fosters loyalty and citizenship behaviors (Ellemers, Kortekaas, & Ouwerkerk, 1999; van der Vegt & Janssen, 2003). Individual affection resulting from identification leads a person to maintain a positive trusting relationship with other in-group members, and therefore elevates his/her activeness of knowledge sharing within the particular team. Indeed, the emotional attachment to a collective supplies the motivational force that leads individuals to collective actions or the readiness to engage in interaction (Bergami & Bagozzi, 2000, p. 563). On the contrary, the arduous relationships, in which situated people feel emotionally laborious and distant to a social group not only suppress their motivations to contribute knowledge but also freeze their motives of learning (Szulanski, 1996). The aforementioned justifies the following hypothesis: Hypothesis 5. An individual’s affective commitment to the belonging team will enhance his/her knowledge sharing in the team.

784

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

Team cooperative norms and knowledge sharing. Team cooperative norms represent a shared value of cooperation among team members. Cooperative norms in a team usually include the willingness to value and respond to diversity, openness to critical thoughts, and the expectation of reciprocity and cooperation (Leonard-Barton, 1995). Team norms guide individual behavior by not only defining what is considered to be appropriate and what should be avoided, but also by providing an organized, interpretable set of informational cues that creates order for individual members. The shared value of cooperation in the team will lead to salient subjective norms regarding cooperation, i.e. an individual’s perception of the expectation of other members in respect to the knowledge sharing with one another (Ajzen & Fishbein, 1980; Bock, Zmud, Kim, & Lee, 2005). The reciprocity embedded in cooperative norms provides team members with some assurance that their knowledge sharing could be rewarded from someone else in the long run, although such a reward may not be immediate and straightforward (Blau, 1964). The aforementioned leads to the hypothesis below: Hypothesis 6. The stronger cooperative norms within a team will increase the nested individuals’ knowledge sharing in the team. 3. Methodology 3.1. Data collection To evaluate the research model, the research team conducted a survey among 9 Chinese organizations. The survey instruments (shown in Appendix A), originally developed in English, were translated into Chinese using Brislin’s (1986) conventional backtranslation method. The main study was facilitated by senior managers of the nine organizations surveyed. To make the respondents feel free to provide their network data, the research team assured them that their responses would not be shared with their supervisors and were only for academic usage. We collected 473 individual observations nested in 65 work teams. These teams engaged in knowledge-intensive work. They were from multiple industries, including engineering and design, software development, telecommunication services and information services. Following the practice of prior research on network properties of organizational teams (e.g. Oh et al., 2004), 18 teams, who had less than an 80% group response rate on the questions about network ties, were excluded in order to improve the reliability of network data solicited from the team members. The final sample was reduced to 343 team members from 47 work teams with an average team response rate of 94.2%. The team size ranged from 3 to 21 members. 68.3% of the respondents were male and 31.7% were female. Their mean age was 35.5 years (s.d. = 10.8) and the mean of their job tenure was 10.3 years (s.d. = 9.2). 3.2. Measures Network data and indices. The network data was collected using the modified ego-centric approach (Wasserman & Faust, 1994), which was also used by several other studies (e.g. Hansen et al., 2005; Obstfeld, 2005). Each team member was asked a series of questions, in which he/she had to list up to 20 names of people in the organization. Four name generator questions were adapted from Obstfeld (2005). The resulting roster of contacts for each work team was the broad network of the team, including intra-team ties and external ties. Based on the solicited networks by team members, each team member’s betweenness centrality in the corresponding network and the network density of each network were calculated. These indices were computed by the widely used UCINET 6 software (Borgatti, Everett, & Freeman, 2002).

Structural Capital. The individual level structural capital is indicated by betweenness centrality. According to Wasserman and Faust (1994), Freeman’s standardized betweenness centrality was used to measure the extent to which each individual member occupied a structurally advantageous position, connecting otherwise unconnected others in the broad team network. The networks were treated as directed ones, taking both egos’ and alters’ evaluations into account. We used network density to measure the team level structural capital. The network density describes the overall level of interaction of various kinds of relations reported by team members. We computed the density of the broad team network for each work team, as the number of existing relations divided by the number of all possible asymmetric relations (Wasserman & Faust, 1994). Cognitive capital. At the individual level, four items adapted from Ko, Kirsch, and King (2005) and Nelson and Cooprider (1996) were used to measure the individual perception of their shared cognition with team members. At the team level, the conceptualization and operationalization of cognition commonality is based on the dispersion-composition model in multilevel research that assumes a certain degree of disparity of individuals’ mental models in each team (Chan, 1998). Thus, the cognition commonality in a team was measured by calculating James’s (1984) interrater agreement index (Rwg ) of team members’ perceptions of the shared cognition between themselves and other members in the same team. The mean of Rwg for team shared cognition was 0.93, ranging from 0.79 to 1. Relational capital. At the individual level, five items were adapted from previous studies to measure individuals’ affective commitment to the team (Van der Vegt & Bunderson, 2005; Wasko & Faraj, 2005). The team level relational capital is represented by team cooperative norms. Six items adapted from Bock et al. (2005) and Kankanhalli, Tan, and Wei (2005) were used to measure the cooperative norms. A degree of consensus among individual members’ ratings to their team environment was expected for assessing the cooperative norms. We followed the direct consensus composition model to measure team cooperative norms by aggregating individuals’ ratings to the team level (Chan, 1998). Knowledge sharing. Knowledge sharing in this study is defined as team members providing and receiving knowledge with other members within the same work team through multiple channels. We distinguish explicit knowledge sharing from tacit knowledge sharing based on the content of knowledge. The measures for the two dimensions of knowledge sharing were adapted from Bock et al. (2005). Specifically, explicit knowledge sharing was measured by the sharing of work reports, manuals, methodologies, etc., and tacit knowledge sharing was measured by the sharing of know-how, know-why, know-whom or know-where, work experiences and expertise from education or training. Control variables. Previous research shows that team size (e.g. Oh et al., 2004; Reagans & McEvily, 2003) and physical distance/proximity (e.g. Darr & Kurtzberg, 2000) of a group of people has influence on their knowledge sharing behavior. We asked the team leader to indicate the team size and physical distance (1 = co-locate in the same office; 2 = disperse across a work block; 3 = disperse across a city; 4 = disperse across the country).

3.3. Data analysis The research model is nested with individual knowledge sharing behavior, and social capital spanning the individual and team levels. The data are also nested, i.e. individuals are nested in work teams. Thus, the Hierarchical Linear Modeling (HLM6) technique was adopted to examine relationships across multiple levels (Bryk & Raudenbus, 1992). The effects of social capital on explicit knowledge sharing and tacit knowledge sharing were tested separately.

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

785

Table 1 Descriptive statistics and correlations. Variables

Mean

s.d.

1

2

3

4

Individual level (n = 343) 1. Explicit knowledge sharing 2. Tacit knowledge sharing 3. Betweenness centrality 4. Shared cognition 5. Affective commitment

5

4.34 4.61 0.04 5.01 5.56

1.19 1.12 0.08 0.93 1.04

0.86 .66*** .12* .45*** .36***

0.90 .05 .48*** .52***

– .13* .10†

0.89 .51***

0.95

Team level (n = 47) 1. Team size 2. Physical distance 3. Network density 4. Cognition commonality 5. Cooperative norms

8.02 1.53 0.13 0.93 5.28

4.23 0.58 0.01 0.06 0.63

– −.15 −.18 −.13 .15

– −.17 .34* .07

– −.12 −.06

– .33*

0.93

Note. Bold values on the diagonal are composite reliability of latent variables. * p < 0.05, two-tailed tests. *** p < 0.001, two-tailed tests. † p < 0.10, two-tailed tests.

4. Results and discussion Before HLM tests, the descriptive statistics were calculated, and the reliability, convergent and discriminant validities for the latent variables were assessed by using Confirmatory Factor Analysis (CFA). Table 1 shows the descriptive statistics and bivariate correlations of variables at individual and team levels. The result of CFA indicated a goodness of fit of the measurement model (2 = 606.17, df = 179, GFI (goodness of fit index) = 0.84, CFI (comparative fit index) = 0.915, RMSEA (root mean square error of approximation) = 0.08). The composite reliability scores of the latent variables ranged from 0.86 to 0.93, exceeding the recommended threshold value of 0.70 (Nunnally, 1978) and thus indicating adequate reliability and convergent validity of the measures for those latent variables. Furthermore, a series of pairwised 2 difference tests were conducted to assess the discriminant validity. The significant 2 differences (range: 5.22–41.55, p < 0.001) between the unconstrained pair models (the pair variables freely correlated) and that of the constrained models (covariance between the pair variables set equal to 1) indicated discriminant validities of the measurements (Bagozzi, Yi, & Phillips, 1991).

centering approach is recommended when the focused multilevel effects are on an incremental perspective. This centering approach facilitates the interpretation of the HLM results and ensures that the individual level effects are controlled for testing of the incremental effects of group level variables. It also lessens multicollinearity in group level estimation by reducing the correlation between the group level intercept and slope estimates (Hofmann & Gavin, 1998; Raudenbus, 1989). Furthermore, to manifest the influences of team social capital on individuals’ knowledge sharing, the effects of other contextual features, such as team size and physical distance, were controlled during the analyses. Table 2 summarizes the results of the HLM analyses. The results demonstrate that the overall social capital held by individual members has significant impact on their knowledge sharing within the teams, explaining 28.12% and 41.76% of the

Table 2 Hierarchical linear modeling results for team member knowledge sharing. Levels and variablesa

4.1. HLM null models HLM null models were run separately for the two individual level dependent variables of interest. Resulting ICC[1] values and associated chi-square tests revealed that 7.6% (2 = 74.09, df = 46, p = 0.006) and 18.2% (2 = 124.4, df = 46, p < 0.001) of the variance in team member explicit knowledge sharing and tacit knowledge sharing resided among teams, respectively. Furthermore, regarding the sampling teams being nested in 9 organizations, another two HLM null models for the dependent variables were run at the organizational level. The variances existing across organizations were less than one percentage, indicating that the major between-groups variances were derived from the difference across teams, instead of across organizations. 4.2. HLM results To test the different effects of multilevel social capital on individual explicit and tacit knowledge sharing, the random coefficient models were run respectively. Only level-1 predictors tested the effects of individual social capital on their knowledge sharing. Intercepts-as-outcomes models added level-2 predictors to test the effects of team social capital on individual knowledge sharing. According to Hofmann and Gavin (1998), both level-1 and level-2 variables were grand-mean centered for HLM analyses. This mean

Level 1 predictors Intercept (Betweenness centrality)2 Betweenness centrality Shared cognition Affective commitment Level 2 predictors Team size Physical distance (Network density)2 Network density Cognitive commonality Cooperative norm Devianceb 2 Rwithin-group 2 Rbetween-groups

2 c Rtotal

Explicit knowledge sharing

Tacit knowledge sharing

M1a

M2a

M1b

4.34*** −10.82* 4.75* 0.46*** 0.21***

4.37*** 4.64*** −12.82* −12.37* 4.93** 4.38* 0.43*** 0.32*** 0.19*** 0.43***

4.66*** −11.78* 3.96* 0.32*** 0.41***

−0.04** 0.15 −8.90† 3.16† 0.49 0.15 977.01

−0.04* 0.22* −8.50* 2.63 −1.30† 0.26** 849.38

988.99 28.12%

863.05 41.76%

M2b

89.50%

51.49%

32.76%

43.53%

a Team members n = 343; Teams n = 47. All models are grand-mean centered. Entries are estimations of the fixed effects (s) with robust standard errors. The italics are control variables. b Deviance is a measure of model fit; the smaller the deviance is, the better the model fits. Deviance = −2 × log likelihood of the full maximum likelihood estimate. 2 2 2 c Rtotal = Rwithin-group × (1 − ICC[1]) + Rbetween-groups × ICC[1]. * ** *** †

p < 0.05. p < 0.01. p < 0.001. p < 0.10.

786

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

within-group variance of explicit and tacit knowledge sharing, respectively. As hypothesized, the coefficient of the square term of Betweenness Centrality on explicit knowledge sharing ( = −10.82, p < 0.05) and tacit knowledge sharing ( = −12.37, p < 0.05) are significantly negative, which means a team member’s betweenness centrality posits an inverted U-shaped influence on individuals’ sharing of their explicit and tacit knowledge within teams; thus, Hypothesis 1 is supported. The individual members’ perception of the shared cognition with other members regarding the team tasks, values, and problem-solving methods are demonstrated as supplying strong cognitive attraction for the within-group sharing of their explicit ( = 0.46, p < 0.001) and tacit knowledge ( = 0.32, p < 0.001); thus, Hypothesis 3 is also supported. As an emotional attachment to a particular team, individual affective commitment is shown to be significantly, positively associated with individuals’ sharing of explicit knowledge ( = 0.21, p < 0.001) and tacit knowledge sharing ( = 0.43, p < 0.001), showing support for Hypothesis 5. It is interesting to find that individuals’ perceived shared cognition over affective commitment exhibits a stronger influence on their sharing of explicit knowledge, whereas individuals’ affective commitment over the perception of shared cognition presents a stronger impact on their tacit knowledge sharing. The intercepts-as-outcomes models for the two types of knowledge sharing reveal that the team social capital is more likely to affect individuals’ tacit knowledge sharing than their explicit knowledge sharing. As shown in Table 2, level-2 predictors explain 89.50% of available 7.2% of the between-groups variance in explicit knowledge sharing, and 51.49% of the available 18.32% of betweengroup variance in tacit knowledge sharing. The team network density has a curvilinear relationship with individuals’ sharing of explicit knowledge ( = −8.90, p < 0.1) and tacit knowledge ( = −8.50, p < 0.05), providing support for Hypothesis 2. Considering the level of significance, we found that individuals’ tacit knowledge sharing within groups required more optimal network configuration, i.e. a moderately dense network, compared with their explicit knowledge sharing. Team cognitive capital, indicated by the commonality and sharedness of team members’ cognitions, is shown to have an insignificant impact on individuals’ explicit knowledge sharing, and surprisingly exhibits a negative, although marginally significant, impact on the tacit knowledge sharing within teams. There is thus no convincing evidence to support Hypothesis 4. With regard to the negative effect, two plausible reasons may explain such a result. First, tacit knowledge sharing could be regarded as the phenomena of “blind or no-look pass basketball” (Cannon-Bowers & Eduardo, 2001). The team cognitive commonality helps individual members to coordinate in an implicit while effective way, and therefore the explicit communication becomes unnecessary. Second, team cognitive commonality to some extent presents the identical knowledge structures among team members. Such knowledge redundancy would hamper knowledge sharing from each other. Hence, team cognitive capital at a higher level has non-significant effect on individuals’ knowledge sharing, and may even decrease the their sharing of tacit knowledge in teams. Team cooperative norms exhibit a significant, positive impact on individual members’ tacit knowledge sharing ( = 0.26, p < 0.01), but present an insignificant relationship with their explicit knowledge sharing ( = 0.15, p = 0.161). Although the evidence only provides partial support for Hypothesis 6, it reveals an interesting finding. Plausibly, sharing explicit knowledge, such as work reports, manuals and progress reports, is a base line for the teams to complete team tasks; however, the sharing of the personally held tacit knowledge cannot be formalized and routinized. Rather, individuals’ tacit knowledge sharing behavior would be guided and

motivated by social norms. These norms could facilitate individuals’ sense-making of the cooperative environment and the expectations of sharing from others. They could also reduce hidden competition, owing to the tacit knowledge sharing per se. As for the control variables, the results demonstrate the negative relationships of team size and knowledge sharing, regardless of the type of knowledge. In larger teams, individuals may perceive less team cohesiveness and less similarity with other members. The effect of physical distance was also checked. As shown in Table 1, the distance of participating teams was short (mean = 1.53, s.d. = 0.58, range = [1–3]). The physical distance has an insignificant relationship with individuals’ explicit knowledge sharing, while it is significantly positively associated with individuals’ tacit knowledge sharing. Thus, a short distance is sometimes not advantageous for individuals to share knowledge. 4.3. Discussion Differentiated from previous studies on social capital at single levels, this study applies a multilevel approach to fragment social capital into individual and team levels to examine their impacts on individual knowledge sharing behavior within work teams. The multilevel analyses reveal that social capital at different levels exerts distinct influences on individuals’ sharing of their explicit and tacit knowledge. The results confirm the importance of individually held social capital for motivating individuals’ engagement in social actions in general, and knowledge sharing in particular. More importantly, this study demonstrates that the cross-level effects of team social capital at a higher level on individual knowledge sharing behavior are not the same, and the effects differ in the different types of knowledge. Team social capital is more important for promoting individuals’ sharing of tacit knowledge than the sharing of explicit knowledge in teams. The research team used social network analysis to quantify the structural capital at the individual level and the team level with betweenness centrality and network density, respectively. The results illustrate that the structural capital at both levels has an inverted U-shaped relationship with team members’ knowledge sharing. Such results provide strong evidence of the notion of optimal network configuration. Oh et al. (2006) postulate that an optimal network is a network with a moderate level of network density, with individuals still retaining various bridging ties linking internal people to external sources. However, this study argues that the optimal social network configuration is founded on the individuals’ structural equivalence and the overall network balance. There is a trade-off between the closure and brokerage network (Reagans & McEvily, 2008). An emerging team network with an excessive individual betweenness centrality will damage the internal cohesion, while a network with an excessive network density will lead to knowledge redundancy and infertility. Both of these extreme network configurations will hamper an individual’s engagement in knowledge sharing in their teams. Instead, individual members who possess positions with moderate betweenness centralities can receive new knowledge from external sources while freeing themselves from the information traffic jam during the knowledge transferring, and thus can share more knowledge. Also, the overall network with a moderate density will reduce the in-group bias, while allowing individual members to maintain a certain degree of internal cohesion, therefore facilitating knowledge sharing in teams. This study distinguishes tacit knowledge sharing from explicit knowledge sharing and reveals that social capital at different levels plays differential roles in motivating individuals’ sharing of different types of knowledge. Individually held social capital supplies the necessary motives for team members to engage in both explicit and tacit knowledge sharing within teams. In spite of the difference in

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

magnitudes, their effects are substantial, confirming the findings in previous research (e.g. Van den Bossche, Gijselaers, Segers, & Kirschner, 2007; Van der Vegt & Bunderson, 2005). However, team social capital may not always exhibit the additive effects on individual knowledge sharing behavior in teams, depending on the content of knowledge. Team social capital is more likely to influence individuals’ tacit knowledge sharing that requires team cohesiveness and cooperative norms to reduce the uncertainty and competitiveness. Also, tacit knowledge sharing is more likely to occur when there is a moderate level of both knowledge overlap and diversity in team members’ cognitive structures. As for the explicit knowledge sharing, individual members still favor the condition with an optimal network configuration, but they may not be so sensitive to the presence of cognition commonality and cooperative norms in the teams.

787

homophily that is based on the perceived similarity facilitates knowledge sharing between individuals and within clusters; on the other hand, such homophily also functions as a barrier to knowledge sharing because it can restrict the acquisition of new knowledge and may instigate entry barriers to those who do not share similar characteristics. The results enrich this notion, revealing that the cognitive structure in different levels requires different levels of homogeneity and the sequent homophily. At the individual level, team members need to psychologically perceive the cognitive similarity, which seeds motivations of knowledge sharing. A lack of common knowledge is likely to frustrate attempts to share knowledge (Reagans & McEvily, 2003). However, the factual high convergence of cognitive structures at the team level may not be a good signal for individual knowledge sharing within the team and may be even worse for the knowledge sharing in the organization. 5.2. Managerial implications

5. Implications, limitations, and future research 5.1. Theoretical implications Social capital does not merely belong to individuals in a social network, but also to the network as a whole. The conceptualization of social capital at multi-levels with empirical support to their effects on intra-team knowledge sharing constitutes the most important theoretical contributions to social capital theory. As work teams have a multilevel nature – individual members nested in teams – single level research that ignores multilevel nested structures will lead to erroneous conclusions (Klein, Dansereau, & Hall, 1994). The research team’s multilevel research, which links individual and team factors, fills the gap present in prior single level studies. The findings that individual social capital and team social capital play different roles in elevating team members’ sharing of different types of knowledge offer precise understanding of the influences of social capital on individuals’ collective behavior. Second, this study sheds light on the appropriateness of an optimal social network configuration for collective actions, contributing to the social network research. A network configuration, in which individuals stand on positions with a moderate betweenness centrality while the overall team network is moderately dense, is able to provide a better social environment for individuals to behave collectively and share their knowledge. Indeed, there is a trade-off between the individual networks and the team network. As with prior social network research (e.g. Burt, 1992; Granovetter, 1973), this study acknowledges the value of bridging ties that link otherwise unconnected people, especially those bridging ties that link external sources with internal members. Meanwhile, this study further pinpoints the importance of balancing the proportion of connections with internal and external ties. Such a balance is not only beneficial for individuals themselves but also for the team as a whole. The optimal social network configuration view, together with the multilevel conceptualization of the individual network and team network in this study, indicates the trade-off between the structural holes theory (Burt, 1992) and the closure view of social network (1990). In this sense, this study provides a way to accommodate the conflict of these propositions in social capital and network research. Despite the balance in social network, this study also teases out the noteworthy balance between cognitive similarity and diversity, pointing toward the third implication. The results demonstrate that individuals’ perceptions of shared cognition with others will facilitate knowledge sharing. A high degree of cognitive commonality and sharedness in a team might not be helpful to individuals’ knowledge sharing, and sometimes it might even impede tacit knowledge sharing. This is consistent with Makela et al.’s (2007) notion of “paradox of homophily”: on the one hand, interpersonal

Practically, the results in this study entail important implications for organizational teams, especially for those that are knowledge-intensive and require more individual members’ involvement in knowledge sharing. First, team leaders or managers can assess the health of their broad team network through the optimal network view. By checking the network density and individuals’ bridging ties, they can advocate an appropriate networking strategy for team members. Team leaders can encourage active internal interactions to foster the network cohesion. A extremely tight-knit team network is not beneficial for individuals’ sharing new knowledge, while a loose-knit network weakens team members’ motivation of knowledge sharing. Team leaders can also encourage more members to engage in boundary-spanning rather than select a few members to act as gatekeeper. This not only helps the team to reduce the risk of knowledge redundancy and maintain continuous rich knowledge sharing in their teams but also helps to avoid potential opportunistic behavior due to an unbalanced position and power in controlling knowledge flow in the network. Thus, the most important takeaway for team leaders is to ensure that the whole team network has an optimal degree of introversion for internal cohesion, as well as extraversion for new knowledge importation. Excessive high or low degrees of either may not benefit, but impair the team effectiveness. Second, team leaders or managers should also pay attention to the human factors in order to motivate individuals for knowledge sharing, especially for tacit knowledge sharing. The pure network structure is not the only reason why individuals behave collectively. More often, the individual perception of sharedness or commonality in cognition structures and the affective commitment supply the immediate motivation and ability for team members to share knowledge. Soft managerial skills, such as cultivating individuals’ cognitive and affective identification with the teams, creating cooperative environments and maintaining desirable norms, controlling the balance of knowledge similarity and diversity among team members, and so forth, are actionable and effective in promoting individuals’ tacit knowledge sharing. Thus, these findings provide new insight for team design and management with respect to prioritizing social resources on enhancing different types of knowledge sharing behavior. 5.3. Limitations and future research First, the research team is aware that the organizational and/or industrial factors that were not accounted for in the current study could have some impact on individual behavior in the contexts. The nested structure of a team points to the fact that individuals reside in a team, the teams reside in an organization, and even the organizations reside in an industry, or a particular society. The team

788

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

properties have the most direct impact on individual behavior. Furthermore, the variances of the latent variables at the organizational level were checked, and the results show less than one percentage of variances across the nine organizations that shared some commonality. This indicates the appropriateness of constraining the social capital to the team level at the upper side for analysis in the sample. Regarding the structural and cultural heterogeneities in different organizations and industries, further investigation involving samples with organizational and industrial varieties will be explored in future research. Second, the investigation illustrates that multi-level social capitals can have similar and different roles on individuals’ sharing of explicit and tacit knowledge in work teams; however, such results are on the basis of post hoc analyses. Our findings partially confirm the tenets that social embeddedness has similar or stronger effects on tacit versus explicit knowledge transfer in the inter-organizational context (Dhanaraj, Lyles, Steensma, & Tihanyi, 2004). Reychav and Weisberg (2009) further examine how tacit knowledge sharing influences explicit knowledge sharing for enhancing the relationships between internal employees and external customers. Thus, future research can investigate the impacts of multilevel social networks on explicit and tacit knowledge sharing from a contingency perspective and find out the conditions by which the determinants have similar or differential effects on different types of knowledge sharing. Also, future research can investigate the interaction of team-level and individual-level social capitals for influencing team members’ knowledge sharing behavior, regarding the fact that individuals simultaneously experience multi-level social capitals.

Basic Research Funds in Renmin University of China from the Central Government (No. 12XNLF04), the Fundamental Research Funds for the Central Universities (No. YWF-13-T-RSC-032), and the Hong Kong Scholar Program.

Appendix A. A summary of measurements Constructs

Measures

References

Knowledge sharing behavior (1 = never; 7 = very often)

How often you shared the following explicit and tacit knowledge with your group members in the past year 1. work reports 2. manuals, methodologies and models 3. work experience or know-how from work 4. contextual knowledge or know-why from work 5. expertise from the education or training 6. know-where or know-whom

Bock et al. (2005), Cumming (2004)

Shared cognition (1 = to a little extent; 7 = to a great extent)

To which extent do you perceived with shared understanding between you and other members in general 1. I agree on what’s important to the work with my team members 2. My team members and I have very similar prior experience of dealing with the confronting tasks 3. My team members and I solve problems in a similar way 4. My team members and I understand each other when we talk

Ko et al. (2005), Nelson and Cooprider (1996)

Affective commitment (1 = to a little extent; 7 = to a great extent)

To which extent do you agree with the following statements 1. I really care about the fate of my team 2. I feel a great deal of loyalty to my team 3. I feel emotionally attached to my team 4. I feel a strong sense of belonging to my team 5. I feel as if the team’s problems are my own

Van der Vegt and Bunderson (2005), Wasko and Faraj (2005)

Cooperative norms (1 = strongly disagree; 7 = strongly agree)

To which extent do you agree with the following statements that describe the work environment of your team 1. openness to freely exchange information 2. openness to conflicting views 3. willingness to value and respond to diversity 4. cooperation 5. collaboration 6. teamwork

Bock et al. (2005), Kankanhalli et al. (2005)

Name generator questions

Looking back over the last year, 1. to whom they turned to for advice 2. with whom they communicated to get work done 3. with whom they discussed important matters 4. who had been influential in getting their work approved

Obstfeld (2005)

6. Conclusions This study, using a multilevel methodology, examines the effects of individually held social capital and team social capital on an individual’s explicit and tacit knowledge sharing behavior in knowledge-intensive work teams. The empirical results give evidence for the distinct role of social capital between levels: individual social capital provides abilities and motives for individual knowledge sharing, while team social capital generates the top-down influences to adjust for individual behavior. The results demonstrate that both individual and team structural capital have an inverted U-shape relationship with the individual explicit and tacit knowledge sharing within the team. This illuminates the appropriateness of an optimal social network configuration for team networking. This study advises management to consider this optimal social network view to their knowledge-intensive work teams, i.e. being open to external sources, while simultaneously being careful to maintain the necessary network density of the teams. The results also underscore the strong influence of individually perceived shared cognition, affective commitment, and team norms on team members’ knowledge sharing. This study suggests that management pay more attention to human cognition and affection, such as team members’ perceptions on peers, the team identity, and the work environment. Humanized management to nurture favored norms would evoke human intrinsic motivation of knowledge sharing, especially for tacit knowledge sharing. Acknowledgements The authors especially thank the editors and the four anonymous reviewers for their constructive comments. The research is partially supported by the National Science Foundation of China (No. 71201165, 71101005), the Research Fund for the Doctoral Program of Higher Education of China (No. 20111102120022), the Aeronautical Science Foundation of China (No. 2012ZG51074), the

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

References Adler, P. S., & Kwon, S.-W. (2002). Social capital: Prospects for a new concept. Academy of Management Review, 27(1), 17–40. Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice Hall. Alderfer, C. P. (1976). Boundary relations and organizational diagnosis. In M. Meltzer, & F. Wickert (Eds.), Humanizing organizational behavior (pp. 142–175). Spring field, IL: Charles Thomas. Argote, L., McEvily, B., & Reagans, R. (2003). Managing knowledge in organizations: An integrative framework and review of emerging themes. Management Science, 49(4), 571–582. Bagozzi, R. P., Yi, Y., & Phillips, L. W. (1991). Assessing construct validity in organizational research. Administrative Science Quarterly, 36(3), 421–458. 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(4), 555–577. Blau, P. M. (1964). Exchange and power in social life. New York: John Wiley. Bock, G.-W., Zmud, R. W., Kim, Y.-G., & Lee, J.-N. (2005). Behavioral intention formation in knowledge sharing: Examining the roles of extrinsic motivators, social-psychological forces, and organizational climate. MIS Quarterly, 29(1), 87–111. Borgatti, S. P., & Cross, R. (2003). A relational view of information seeking and learning in social networks. Management Science, 49(4), 432–445. Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2002). UCINET 6 for windows. Columbia, SC: Analytic Technologies. Brass, D. J., Galaskiewicz, J., Greve, H. R., & Tsai, W. (2004). Taking stock of networks and organizations: A multilevel perspective. Academy of Management Journal, 47(6), 795–817. Brislin, R. (1986). The wording and translation of research instruments. In W. Lonner, & J. Berry (Eds.), Field methods in cross-cultural research (pp. 137–164). Beverly Hills, CA: Sage Publications. Bryk, A. S., & Raudenbus, S. W. (1992). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage. Burt, R. S. (1992). Structural holes: The social structure of competition. Cambridge, MA: Harvard University Press. Cannon-Bowers, J. A., & Eduardo, S. (2001). Reflections on shared cognition. Journal of Organizational Behavior, 22(Special issue), 195–202. Chan, D. (1998). Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83(2), 234–246. Coleman, J. S. (1990). Foundation of social theory. Cambridge, MA: Belknap/Harvard University Press. Cumming, J. N. (2004). Work groups, structural diversity, and knowledge sharing in a global organization. Management Science, 50(3), 352–364. Darr, E. D., & Kurtzberg, T. R. (2000). An investigation of partner similarity dimensions on knowledge transfer. Organizational Behavior and Human Decision Processes, 82(1), 28–44. Dhanaraj, C., Lyles, R. A., Steensma, H. K., & Tihanyi, L. (2004). Managing tacit and explicit knowledge transfer in IJVs: The role of relational embeddedness and the impact on performance. Journal of International Business Studies, 35, 428–442. Ellemers, N., Kortekaas, P., & Ouwerkerk, J. W. (1999). Self-categorisation, commitment to the group and group self-esteem as related but distinct aspects of social identity. European Journal of Social Psychology, 29(2–3), 371–389. Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–1380. Grant, G. M. (1996). Prospering in dynamically-competitive environments: Organizational capability as knowledge integration. Organization Science, 7(4), 375–387. Hansen, M. T., Mors, M. L., & Lovas, B. (2005). Knowledge sharing in organizations: Multiple networks, multiple phases. Academy of Management Journal, 48(5), 776–793. Hau, Y. S., Kim, B., Lee, H., & Kim, Y. G. (2013). The effects of individual motivations and social capital on employees’ tacit and explicit knowledge sharing intentions. International Journal of Information Management, 33(2), 356–366. Hoegl, M., Parboteeah, K. P., & Munson, C. L. (2003). Team-level antecedents of individuals’ knowledge networks. Decision Sciences, 34(4), 741–770. Hofmann, D. A., & Gavin, M. B. (1998). Centering decisions in hierarchical linear models: Implications for research in organizations. Journal of Management, 24(5), 623–641. James, L. R. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69(1), 85–98. Kang, M., & Kim, Y. G. (2010). A multilevel view on interpersonal knowledge transfer. Journal of the American Society for Information Science and Technology, 61(3), 483–494. Kankanhalli, A., Tan, B. C. Y., & Wei, K. K. (2005). Contributing knowledge to electronic knowledge repositories: An empirical investigation. MIS Quarterly, 29(1), 113–143. Klein, K. J., Dansereau, F., & Hall, R. J. (1994). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19, 105–229. Klimoski, R., & Mohammed, S. (1994). Team mental model: Construct or metaphor? Journal of Management, 20(2), 403–437. Ko, D.-G., Kirsch, L. J., & King, W. R. (2005). Antecedents of knowledge transfer from consultants to clients in enterprise system implementations. MIS Quarterly, 29(1), 59–85.

789

Leonard-Barton, D. (1995). Wellsprings of knowledge: Building and sustaining the source of innovation. Boston: Harvard Business School Press. Makela, K., Kalla, H. K., & Piekkari, R. (2007). Interpersonal similarity as a driver of knowledge sharing within multinational corporations. International Business Review, 16(1), 1–22. Mehra, A., Kilduff, M., & Brass, D. J. (2001). The social networks of high and low self-monitors: Implications for workplace performance. Administrative Science Quarterly, 46(1), 121–146. Mohammed, S., & Dumville, B. C. (2003). Team mental models in a team knowledge framework: Expanding theory and measurement across disciplinary boundaries. Journal of Organizational Behavior, 22(Special issue), 89–106. Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational advantage. Academy of Management Review, 23(2), 242–266. Nelson, K. M., & Cooprider, J. G. (1996). The contribution of shared knowledge to IS group performance. MIS Quarterly, 20(4), 409–432. Nunnally, J. C. (1978). Psychometric theory. New York: McGraw Hill. Obstfeld, D. (2005). Social networks, the tertius lungens orientation, and involvement in innovation. Administrative Science Quarterly, 50(1), 100–130. Oh, H., Chung, M.-H., & Labianca, G. (2004). Group social capital and group effectiveness: The role of informal socializing ties. Academy of Management Journal, 47(6), 860–875. Oh, H., Labianca, G., & Chung, M.-H. (2006). A multilevel model of group social capital. Academy of Management Review, 31(3), 569–582. Perlow, L. A., Gittell, J. H., & Katz, N. (2004). Contextualizing patterns of work group interaction: Toward a nested theory of structuration. Organization Science, 15(5), 520–536. Phelps, C., Heidl, R., & Wadhwa, A. (2012). Knowledge, networks, and knowledge networks: A review and research agenda. Journal of Management, 38(4), 1115–1166. Podolny, J. M., & Baron, J. N. (1997). Resources and relationships: Social networks and mobility in the workplace. American Sociological Review, 62(5), 673–693. Putnam, R. D. (1993). The prosperous community: Social capital and public life. The American Prospect, http://www.prospect.org/print/v4/13/putnamr.html Raudenbus, S. W. (1989). “Centering” predictors in multilevel analysis: Choices and consequences. Multilevel Modeling Newsletter, 1, 10–12. Reagans, R., & McEvily, B. (2003). Network structure and knowledge transfer: The effects of cohesion and range. Administrative Science Quarterly, 48(2), 240–267. Reagans, R., & McEvily, B. (2008). Contradictory or compatible? Reconsidering the “trade-off” between brokerage and closure on knowledge sharing. Network Strategy Advances in Strategic Management, 25, 275–313. Reychav, I., & Weisberg, J. (2009). Going beyond technology: Knowledge sharing as a tool for enhancing customer-oriented attitudes. International Journal of Information Management, 29, 353–361. Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best practice within the firm. Strategic Management Journal, 17, 27–43. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987). Rediscovering the social group: A self-categorization theory. Oxford/New York: Blackwell. Van den Bossche, P., Gijselaers, W. H., Segers, M., & Kirschner, P. A. (2007). Social and cognitive factors driving teamwork in collaborative learning environments. Small Group Research, 37(5), 490–521. Van der Vegt, G. S., & Bunderson, J. S. (2005). Learning and performance in multidisciplinary teams: The importance of collective team identification. Academy of Management Journal, 48(3), 532–547. van der Vegt, G. S., & Janssen, O. (2003). Joint impact of interdependence and group diverisity on innovation. Journal of Management, 29(5), 729–751. Wasko, M. M., & Faraj, S. (2005). Why should I share? Examining social capital and knowledge contribution in electronic networks of practice. MIS Quarterly, 29(1), 35–57. Wasserman, S., & Faust, K. (1994). Social network analysis. New York: Cambridge University Press. Wei, J., Zheng, W., & Zhang, M. (2011). Social capital and knowledge transfer: A multi-level analysis. Human Relations, 64(11), 1401–1423. Xiao, Z. X., & Tsui, A. S. (2007). When brokers may not work: The cultural contingency of social capital in Chinese high-tech firms. Administrative Science Quarterly, 52(1), 1–31. Yang, S. C., & Farn, C. K. (2009). Social capital, behavioural control, and tacit knowledge sharing: A multi-informant design. International Journal of Information Management, 29, 210–218. Zagenczyk, T. J., Scott, K. D., Gibney, R., Murrell, A. J., & Thatcher, J. B. (2010). Social influence and perceived organizational support: A social networks analysis. Orgnizational Behavior and Human Decision Processes, 111(2), 127–138. Zhu, B., & Watts, S. A. (2010). Visualization of network concepts: The impact of working memory capacity differences. Information Systems Research, 21(2), 327–344. Yan Yu is an Assistant Professor in Information School at Renmin University of China. She received her Ph.D in information systems from City University of Hong Kong and received her master degree and dual bachelor degrees from Peking University. Her research interests include IT-enabled Knowledge Management, Enterprise 2.0 and innovation, and IT strategy. She has published her work in several international journals including Decision Support Systems, Journal of Knowledge Management, IT & Behaviour, Computers & Education, etc. She is also active in the prestigious IS conferences such as ICIS, ECIS, AMCIS, PACIS and HICSS. Jin-Xing Hao is a “Hong Kong Scholar” at the School of Hotel and Tourism Management, Hong Kong Polytechnic University and an Assistant Professor at the School

790

Y. Yu et al. / International Journal of Information Management 33 (2013) 780–790

of Economics and Management, Beihang University. He received his PhD from City University of Hong Kong. His research interests include knowledge management, business intelligence and tourism technology management. He has published articles in journals such as Journal of American Society for Information Science and Technology, Decision Support Systems, IEEE Transactions on Knowledge and Data Engineering, Journal of Information Science and others. Xiao-ying Dong is an Associate Professor in Guanghua School of Management at Peking University. She received her Ph.D from Peking University and academically visited Harvard University, University of Pittsburgh, Australian National University. Her research interests include Knowledge Management Strategy, Organizational Ambidexterity, and Dynamic Capability Building. She has published 5 books and a number of refereed articles in international journals and conferences such as Information Systems Journal, ICIS, ECIS, AMCIS, and PACIS.

Mohamed Khalifa received the degrees of M.A. in Decision sciences and PhD in Information Systems from the Wharton Business School of the University of Pennsylvania. He is currently the Vice President, Academic, at Al Ghurair University. His work experience includes 20 years as a business analyst and senior consultant and 17 years as an academic in the United States, Canada, China, Hong Kong and the UAE. Dr. Khalifa conducted extensive research in the areas innovation adoption, electronic commerce, IT-enabled learning and knowledge management. He has published over 120 research articles. His work appeared in journals such as Journal of Business Research, OMEGA, Communications of the ACM, Journal of the Association of Information Systems, IEEE Transactions on Engineering Management, IEEE Transactions on Systems Man and Cybernetics, Decision Support Systems, Data Base and Information and Management.