بررسی شبکه های اجتماعی در سطح تیم - بررسی ادبیات تجربی

Despite the extensive set of findings on the determinants of team effectiveness, academic understanding of one potentially critical set of determinants, social networks, is limited. This paper is a review and a discussion of the literature on the impact of social networks in small groups such as teams. More specifically, the interest is in the effects of the structural characteristics of the networks on team effectiveness. The review covers various types of small groups: subjects of laboratory studies, student teams, innovation and R&D teams, and other organisational groups. The research findings for each type are reviewed, and the article concludes with a comparison of the variables studied. The progress that has been made is highlighted, suggestions for further research are made, and the key contributions to this area of study are summarised.

The importance of interaction is acknowledged as a basic process in an organisation. It enables the development and maintenance of its goals: humans working together need to constantly find effective ways of creating and maintaining the flow of ideas, information, decisions and tasks. The “human side” also has a clear role in new-product development, technology implementation and technology transfer. Thus, technology managers are among those who are acutely concerned about human connections (Green and Aiman-Smith, 2004), also termed social networks. The literature on networks in general is extensive. Ranging from the social to the organisational and beyond, networks have emerged into a research area that includes and brings together various fields such as organisational theory and behaviour, strategic management, business studies, health-care services and public administration (Provan et al., 2007). In general, academics have made considerable efforts to understand what networks really are and how they develop. Despite the progress there is still a great deal we do not know. The aim of this article is not to reiterate what is already familiar about networks in general, but rather to focus on one particular aspect of network studies that has received only scant attention until recently. The main objective, therefore, is to establish what we know about the impact of social networks on various types of small groups such as teams. To date there has been relatively little research conducted in the area of social contacts and their consequences for the functioning of small groups (Krackhardt, 1990 and Ibarra and Andrews, 1993). Previous studies on social networks have tended to focus on the structural properties of egocentric (e.g., employee) or bounded (e.g., organisational) networks as the unit of analysis (Cummings and Cross, 2003). Furthermore, in numerous cases researchers have concentrated on the factors that influence small group (i.e. team) effectiveness (Kozlovski and Bell, 2003), and social networks have not often been included in the critical set of determinants. Increased competition, shortening life-cycles, increased customer requirements, developing technology and globalisation are often suggested as reasons for the need to innovate and develop the products and services companies bring to the market (Belliveau et al., 2002, Rosenau et al., 1996 and Belliveau et al., 2004 cf. Leenders et al., 2007). The use of small groups such as teams has dramatically expanded in response to these competitive challenges (Manz and Sims, 1993). They have thus become the central building blocks of a modern organisation, and practitioners and academics have increasingly started to stress their importance in achieving organisational success in the current economic climate. Some scholars suggest that the ability to transfer knowledge represents a distinct source of competitive advantage in comparison to other institutional arrangements such as markets (Kogut and Zander, 1992), and the effective transfer of knowledge among individuals is important or even critical in a variety of organisational processes and outcomes, such as the transfer of best practices (Szulanski, 1996), new-product development (Hansen, 1999), and even organisational survival (Baum and Ingram, 1998). Interpersonal social networks are considered to play a central role in this process. Teams could be considered information-processing units: like individuals they encode, store and retrieve information (Brauner and Scholl, 2000). They exchange it through effective interaction and building on the knowledge of others, and can create new knowledge and insights (Moenaert et al., 2000 and Bakker et al., 2006). Consultation and interaction could help them to foresee and possibly rule out potential weaknesses in technical and marketing solutions, for example. In other words, inter-team communication represents the logistics through which knowledge is accessed, transferred, and absorbed into new knowledge, ideas and insights. Developing, refining, testing, selecting and implementing ideas are all dependent on interaction among team members. Furthermore, a higher level of interaction makes cross-fertilisation more likely, thus potentially fostering more and better ideas (West, 1990). It is therefore clear that managers of technological-innovation teams, for example, also need to take care of the human connections (Green and Aiman-Smith, 2004). All this makes it increasingly important to understand the relationship between both intra-team and inter-team social relations and team effectiveness, also for technology managers. More specifically, the aim of this review is to find out what previous research reveals with regard to the following questions. Do social networks of individuals within the small group such as team have an impact on group effectiveness? Do patterns of group-internal social networks affect the effectiveness? Do group-external social networks have an impact? In her search for answers the author reviewed the research on various team-working arrangements, ranging from early laboratory investigations in the 1950s and 1960s to more recent research streams starting from the 1980s (32 studies in total). The variables studied in different types of small groups are compared, the progress that has been made is highlighted, suggestions for future action are given, and the key lessons learned from the review are summarised. A social network in the context of this review is “a set of nodes and the set of ties representing some relationship, or a lack of relationship between the nodes” (Brass et al., 2004, 795) within a small group such as team. Thus, the idea of studying social networks in small groups stems from the underlying concept of the network approach (Wellman, 1988), which describes how the structure of social interaction creates access to specific resources. The question of whether the social-network tradition is based on any real theory or theoretical approach has aroused a great deal of debate among researchers in this field. Others rather see it as an “orientation towards the social world” and “a collection of methods” (Scott, 2000, 27), or “as a theory of social structures” (Degenne and Force, 1999, 12). In this research and in this respect I follow the ideas presented by Kilduff and Tsai: social network theory is not a single entity but rather a collection of theories under one umbrella (Kilduff and Tsai, 2007). The main point is rather simple: it concerns whether the social networks in which people are embedded have an impact on their behaviour. More specifically, people's behaviour depends on their interaction with one another and their relationships with the overall social network. The first task is thus to define a team, types of teams and team effectiveness. The methods and scope of the study are then described in more detail (Section 2), and the criteria for selecting the literature are explained (Section 3). Section 3 also presents the research findings for each type of team, organised in accordance with the variable categories. The paper concludes with a summary of the key findings on what was learned about social networks in small groups such as teams (Section 4) and recommendations for further research (Section 5).

By way of a conclusion I will first compare the different types of teams in terms of the categories of variables under investigation (i.e. measures of team social networks and team effectiveness) and the setting in which the studies were conducted. I will then assess the progress that has been made, and what still needs to be done in studying social networks in teams. Finally, I will summarise the four key lessons learned, and highlight areas for future research. 5.1. The different types of teams compared Table 6 compares the study focus for the different types of teams, classified as before into laboratory experiments, student teams, innovation and R&D teams and other organisational groups, and maintaining the earlier categorisation of social-network variables into individual, team-internal and team-external levels. The outcome categories represent the different dimensions of effectiveness (objective, perceptual, attitudinal and behavioural) investigated for each type of team, thereby providing a starting point for categorising how teams vary in type. Table 6 lists the types of social-network ties investigated in the reviewed studies. The ones most often examined were based on communication, classified as task-related (e.g., “How frequently do you communicate with [name of the person] during the project?”), advice-related (e.g., “Who do you go to for advice on work-related matters?”) and social (e.g., “With whom do you interact socially outside work?”). A variety of other types of ties were also investigated, such as workflow, problem-solving, awareness and preference networks. Another major distinction is between strong and weak network ties, but the studies made scant reference to strength (Shah et al., 2006, student teams; Hansen, 1999, innovation and R&D teams). It would therefore be interesting to investigate this further in that it involves various issues such as reciprocity and intensity. There was some mention of negative ties in the studies (e.g., “Does [name of the person] make it difficult for you to carry out your work?”), which characterise relationships that are generally considered to inhibit (team) performance. The Laboratory studies did not focus on such ties. The lack of studies may be due to the problems inherent in obtaining valid data on negative relationships in the first place. Table 6 also shows that the early and more recent studies on different types of teams differ only slightly in terms of predictor variables. The majority, including the early studies and involving all kinds of small groups, include density (or some variation of it) and centralisation, or alternatively a measure of group partition (see Wasserman and Faust, 1994 for the term) such as structural holes, in the analysis. However, the way these measures are operationalised may vary considerably, as discussed at the beginning of this section. The studies also vary in how they measured team effectiveness. Thus, it seems that there is no commonly accepted meaning of team effectiveness. The early studies used very context-specific measures, including the number of messages sent and the time taken to solve the simulated problems (see e.g., Guetzkow and Simon, 1955 and Heise and Miller, 1951). Those on innovation and R&D teams generally applied objective measures of quantitative research outcomes such as publications and reports, and developmental outcomes such as experimental prototypes and materials. This type of approach may have been driven by the convenience of operational measures. In this context, the dimensions of effectiveness seem, to a large extent, to correspond with managerial information needs. Surprisingly, effectiveness was seldom treated as a multi-dimensional concept, the focus often being only on team performance. This may be due to the fact that collecting enough social-network data for statistical analysis requires a huge effort. The most frequently used measures of team effectiveness were perceptions gained from managers or supervisors, followed by internal team perceptions or perceptions that included both managers and team members. Researchers studying team social networks could consider expanding their data-collection tools beyond the self-reports of managers and team members, and in the context of “other organisational groups” they should complement the use of surveyed performance measures with more objective tools. One of the reasons for the lack of objective measures in this category in the reviewed studies may be the fact that most of the teams were project groups or otherwise short-lived. It is generally considered difficult to execute objective performance measures on project teams because the outputs may extend too far into the future and hence are difficult to assess (Sundstrom et al., 1990). Another option would be to rely more on the evaluations of people other than team leaders and members in order to improve reliability. The combination of objective and subjective measures could be considered, given the fact that the accuracy of the latter relies on the “goodness” of the evaluators - higher-level managers, for example. Interestingly, Smith-Doerr et al. (2004) found that social networks not only had an impact on information access, but also moulded the manager's impressions of the success of innovative projects. Using multiple measures of effectiveness is another option. Effectiveness beyond task performance, including some attitudinal and behavioural aspects, was only measured in the early laboratory studies and in student teams (Lucius and Kuhnert, 1997; see Jehn and Shah, 1997). In the future there should be more focus on attitudinal and behavioural outcomes. Furthermore, only some of the reviewed studies explicitly point out that there are several data-collection phases in project work (see Cummings and Cross, 2003 and Yang and Tang, 2004), or before it is completed (Reagans et al., 2004). However, in general it seems that both network and effectiveness data were collected after the outputs had been produced, and it is thus not certain whether the social networks affected the performance or vice versa. There is thus a need for more longitudinal studies: it is acknowledged that they are challenging and time-consuming, but they are the best way of assessing causality. 5.2. Progress has been made It is clear that considerable effort has been made since the early laboratory studies in the 1950s and 1960s, but there is still research to be done. There have recently been some shifts in emphasis that have enriched our knowledge of team networks by filling in some of the major gaps. I list these shifts here in order to support them and to document them. Following the early “MIT studies” the research starting from the 1980s has progressed from simulated settings towards studies conducted in organisations. Further, as Table 6 shows, there has been another shift from single-level to cross-level analysis, which was probably inspired by the notion that social networks form both from below (individuals within groups/teams) and above (embedded in departments, units and in the larger organisation). There is also a detectable shift in focus from communication networks to the consideration of distinctions such as the type of relationship. The more recent studies covered various types of networks ranging from the expressive and instrumental to the negative. Table 6 gives more details about this variation: this level of detail is often required in order to make theoretical predictions. There has also been a slight shift from the static to the dynamic, although the trend is not yet very strong. Few current studies show an interest in how networks evolve (see Shrader et al., 1989, Kratzer et al., 2005 and Yang and Tang, 2004). It is also clear that the content and the strength of social-network relationships have started to interest current researchers (see e.g., Hansen, 1999). The early studies concentrated more on calculating how many contacts were needed in order to carry out a certain task, for example, and the interest was thus merely in the existence or non-existence of a relationship. This is not solely the case today. 5.2.1. What has been learned about social-network effects in teams? The key findings in this respect are given below, the aim being to draw an initial sketch of the impact of social networks on team effectiveness. The substantive findings are summarised in Fig. 2. It is clear from the review that social networks in teams have consequences for team effectiveness. However, the empirical evidence (number of studies) linking team networks to team effectiveness (e.g., performance) is still fairly moderateand a lot more research needs to be done. There is not yet enough information to draw up detailed criteria for deciding on the type and patterns of relational structures that are beneficial or detrimental, such as whether the existing network needs improvement, or what the optimal form is. The interest has been general, as Fig. 2 shows, and only a few issues have been considered in more detail. Density was the aspect that attracted most attention in all the different kinds of teams investigated. Density has been found to affect team performance positively, especially in student teams. We can only speculate why this is the case, but it may be that the nature of student teams is often essentially about communication. Thus, groups that communicate more are also likely to produce better outputs. There have been fewer studies on the relationship between density and effectiveness in the other types of groups/teams, and the investigations have produced contradictory results. Some of these results may be at least potentially attributable to the different types of networks addressed. For example, Kratzer et al. (2005) found that friendship networks had a positive impact on perceptual performance in innovation teams whereas the impact of friendly networks was inversely u-shaped. In fact, Priem and Rosenstein (2000) suggest that it is positive when alternative explanations appear in that the old explanations can be disputed and future research can build on the new ones. At least in this case it shows that we are dealing with an evolving area of research. Fig. 2 also shows how previous research has focused more on the performance-related outcomes of dense networks, although the few studies that have dealt with other types of outcomes are fairly unanimous in concluding that the impact on team satisfaction, for example, is positive. In general, the reviewed studies have generally succeeded in explaining the variance in objective and perceptional performance, and only to some extent in attitudinal and behavioural outcomes. As Fig. 2, which summarises the substantial results, shows, we seem to be lacking studies on attitudinal and behavioural outcomes in teams (see e.g. Lucius and Kuhnert, 1997 and Jehn and Shah, 1997, for exceptions). It would be useful to develop more multidimensional measures of such outcomes in order to assess the consequences of the social networks in a more fine-grained manner. This could also give a better insight into the relationships between the different social structures and team effectiveness. However, qualitative goals concerning the perceptions of the most important stakeholders such as team managers and top management may be more important sometimes. Furthermore, the time at which team effectiveness is addressed may affect the meaning of effectiveness. This again leads to the question of what dimensions effectiveness consists of. The different types of measures used and their operationalisation in the reviewed studies are summarised in Table 7 with a view to facilitating further research in this area. Naturally, selecting measures just because they have been used previously may not be the best strategy if the aim is to enhance understanding of social-network performance in different types of teams. The notion that organisations are influenced by the environment in which they are embedded is classic in organisational theory, and it is not likely to be challenged. The early studies on team social networks tended to forget the environment in which they were embedded (see the general criticism on network research in Ibarra et al., 2005). However, it seems that some researchers have already taken on the task of examining networks at the boundary of the team and its environment. As Fig. 2 shows, researchers studying innovation and R&D teams in particular have focused on team-external networks. These types of groups are likely to be involved in complex problem-solving activities and would thus need external contacts in order to gain access to a variety of knowledge, for example. External networks have been identified as determinants of objective outcomes in innovation and R&D teams, and were also investigated in student teams and in two organisational groups. A perceptual performance impact was found, which might indicate that it is not just innovation but also work teams, “business as usual”, that could benefit from external relationships. As Ancona (1990) found, high-performing teams carried out more information-gathering activities than their lower-performing counterparts. 5.3. Areas for future research 1 In general, more research is needed in order to be able to draw stronger causal relationships and conclusions in this field of study (see also Fig. 2). It may also be that the different types of social structures assume more or less importance depending on the type of team. Thus, replicating similar analyses undertaken in other settings may help in building up a general theory.One setting we do not seem to know much about so far is that of the virtual team: only one study that clearly focused on virtual student teams (Glückler and Schrott, 2007) was found. There is no doubt, however, that virtual teams are becoming an increasingly common feature of the organisational landscape. In the late 1970s Allen et al. found that physical proximity enhanced information processing. Virtual teams lack physical proximity because their members do not all reside in the same location. Researchers should therefore be encouraged to investigate them in order to extend the theory and to strengthen explanations of social networks as predictors of team effectiveness. From the managerial point of view this type of research could provide further insights into the use of new technologies in improving team effectiveness. 2 Future studies could also address behavioural and attitudinal outcomes instead of focusing on task-related performance, given the lack of such investigations in the sample. One could question whether these different types of team effectiveness have similar or different predictors. Both social and task-related outcomes have been highlighted as significant aspects of team effectiveness (see e.g., Kozlovski and Bell, 2003 for a review). 3 Furthermore, there were only three studies (of the thirty-two) that examined both affective and instrumental ties (DiMaggio, 1992). It would be beneficial to include both types of ties in order to find out which ones really matter in terms of understanding the performance of teams. It was found that advice networks in particular transferred work-related knowledge, but emotional-support networks did not. (Henttonen et al., 2009) Moreover, different types of ties may have different roles during the team-development phase. Affective ties (such as emotional support) are likely to take more time to evolve, for example, and at the beginning the ties are mainly instrumental. Later on the situation may be different (see Kratzer et al., 2005). Furthermore, a clear finding was that, as in the literature on networks in general, only a few (three) of the reviewed studies dealt with negative relationships (Sparrowe et al., 2001, Baldwin et al., 1997 and Yang and Tang, 2004). Researchers might consider conducting more in-depth studies on groups in which there are interpersonal dislikes, conflicts and disputes. This could also involve more thorough investigation into the downsides of social networks on the team level. The results might be surprising. Baldwin et al. (1997), for example, found that even if conflicts were not liked, they played a role in achieving higher team performance. 4 One focal question in this review concerned the impact of an individual's social networks on team effectiveness. It is an area that seems to have attracted only limited attention (only five studies) in various types of teams concentrating on different dimensions of efffectiveness. Given the difficulty in drawing conclusions based on the few studies concerned, I would like to suggest it as an area for further research. What is the role of an individual in the creation of team networks and on team-level effectiveness? What if a key individual in the network does not have enough task knowledge, or blocks the transfer of knowledge to other team members? Alternatively, could the team benefit from having a key individual with access to diverse sources of information and who might even function as a gatekeeper? Moreover, if small groups such as teams consist of diverse people, are certain networks related to certain individuals, such as team leaders? I could find only three studies linking leadership to group effectiveness. They all demonstrate the importance of leader-member exchange by showing that it affects a variety of team outcomes such as customer loyalty (Mehra et al., 2006), group creativity (Kratzer et al., 2008) and group performance (Cummings and Cross, 2003). We could therefore ask to what extent the centrality of a leader helps him/her to make the best decisions. Do the relationships between the leader and the members inhibit the leader or constrain his/her actions so that she/he is not willing to make decisions that have negative consequences for the team members? What are the implications in terms of team effectiveness? 5 Previous research has produced some consistent findings in relation to team-external networks. Further studies could investigate not just the impact of the density of the external ties on team performance, but also the type of portfolio of external ties that facilitates it. Studies on team social networks should also take more account of the context. We need to understand the social relationships in which teams are embedded. I also suggest developing a multi-level perspective locating team networks within the larger context of organisations, and considering the impact of team-external relationships - not only on the intra-organisational level as in previous research but also on the inter-organisational level. The recent theoretical models of group social capital offered by Oh et al. (2006), Goyal and Akhilesh (2007) and Evans and Carson (2005) do not include company-external relationships. In practice, these types of cooperative arrangements between companies are often conducted through teams. Hence, further development would be beneficial, and would provide deeper insights into how the potential of inter-organisational relationships is integrated on the team level. Inherent in this type of enquiry would be the need to see teams as organisationally embedded units of analysis to an even greater extent than we do now. 6 Further research could also investigate whether the different social-network structures interact with each other and thereby affect team effectiveness. For example, as we already have some evidence of the benefits of dense intra-team ties, future research could examine whether the centrality of an individual moderates the effects of the team's social-network density. At least to my knowledge there are no studies investigating this matter on the team level. In particular, researchers investigating innovation and R&D teams could benefit from adding some aspects of group partition (Wasserman and Faust, 1994) to their models given the lack of attention in existing studies. There is also a small stream of research based on the premise that optimal network configurations combine seemingly conflicting elements such as cohesion and range (see, for example, Reagans and Zuckerman, 2001). The optimal configuration of ties may thus be a combination of network density and range (Reagans and Zuckerman, 2001). The results of this research are in line with Obstfeld's (2005) notion that both network structures are needed in order to promote innovation and creativity, not to mention cooperation and coordination. Thus, density facilitates integration of the diverse knowledge and capabilities made accessible by the range. The complementaries and tensions between the different network structures are acknowledged, but studies addressing these types of issue on the team level are, to the author's knowledge, rare. 7 Finally, in reality researchers seem to study teams in static settings even though they may aim to investigate team dynamics (McGrath, 1986). Regardless of the fact that networks are dynamic, the research on team social networks still mainly comprises snapshots of the network dynamics. There are studies that have addressed time and temporal patterns in the context of groups (Worchel, 1994 and McGrath, 1991) from which researchers could benefit. They could, for example, investigate whether social structures differ across small groups such as teams in evolutionary stages. Some of the reviewed studies did control for the impact of the project-work phase ( Kratzer et al., 2005 and Yang and Tang, 2004), but did not explicitly report the performance effects. In general, it seems that the temporal element has received less attention in studies on social networks in teams, with the exception of Shrader et al. (1989). We could also question further whether team-network evolution occurs in predictable ways. Are there specific evolutionary stages? Does a lack of success within the team lead to more dense teams, or is it successful collaboration that does so? Is there a feedback loop between team effectiveness and social network structures? All in all, the set of findings described in this article is promising. It is suggested that intertwining the literature on teams and social networks is a fruitful focus and deserves more research attention. Let us hope that these findings are just the beginning of a long story.