Knowledge about the distribution of information and group decision making: When and why does it work?

https://doi.org/10.1016/j.obhdp.2008.10.003Get rights and content

Abstract

Research has shown that decision-making groups with distributed information perform better when group members know which member is knowledgeable about what. Thus far research has been unable to identify the process responsible for this effect. In the present study, we propose that group members’ task representations mediate the effect of knowledge about the distribution of information on decision performance. Building on this proposition, we also propose that reflection about the task moderates the effect of knowledge about distributed information through its effect on task representations. These hypotheses were put to the test in an experimental study of decision-making groups (N = 125). As predicted, knowledge of distributed information interacted with reflection to affect decision quality. Findings confirmed the proposed mediating role of task representations and information elaboration.

Introduction

Because of groups’ potential to pool and integrate informational resources, decision-making tasks are often allocated to groups rather than to individuals. Although in theory the greater amount of resources held by groups should allow them to make better decisions, numerous studies have documented how groups with information that is distributed over group members do not adequately use this information in decision making. Not only do group members often fail to exchange their informational resources (e.g., Gruenfeld et al., 1996, Stasser and Titus, 1985, Wittenbaum and Stasser, 1996), when they do exchange information they often fail to integrate the information into their actual decision, which may result in lower-quality decisions (Gigone and Hastie, 1993, Scholten et al., 2007, Winquist and Larson, 1998). Accordingly, from a theoretical as well as from an applied perspective it is important to identify factors that influence the extent to which groups elaborate on (i.e., exchange and integrate) distributed information (De Dreu et al., 2008, Hinsz et al., 1997, van Knippenberg et al., 2004).

In the present study, we focus on a factor that is somewhat of an enigma in this respect—knowledge about the distribution of information within the group (i.e., knowledge about who knows what). Research has reliably shown that when group members in groups with distributed information have knowledge about who knows what, they exchange more information and make higher-quality decisions (e.g., Littlepage et al., 1997, Stasser et al., 1995, Stasser et al., 2000, Stewart and Stasser, 1995). At the same time, however, it is unclear why knowledge about who knows what has this effect. Although it makes intuitive sense that knowing who knows what helps groups to use their distributed information, obvious explanations for the effect are not supported by the data and it remains unclear which processes account for it (Stasser et al., 2000). Because we do not know what causes the effect of knowledge about information distribution on group decision quality—a question of theoretical importance—it is also unclear under which circumstances the positive effect of knowledge about information distribution is more likely to obtain—a question of theoretical and practical importance. In the present study, we address both these issues.

To do so, we rely on a recent analysis by van Ginkel and van Knippenberg (2008) that puts group members’ understanding of the decision-making task (their task representations; Newell and Simon, 1972, Tindale et al., 1996) center-stage in groups’ ability to use distributed information. Building on this analysis, we propose that when group members know who knows what, this affects their understanding of the group task and how best to approach the task. More specifically, we propose that knowledge about the distribution of information leads to task representations that emphasize the importance of elaboration of decision-relevant information. These task representations in turn stimulate the elaboration of distributed information and decision-making performance. In short, we propose that group members’ knowledge of who knows what influences decision-making performance because it affects their understanding of the task.

For our understanding of groups’ ability to use their informational resources it is important to identify the process that mediates the effect of knowledge about distributed information. We aimed to further develop the theoretical contribution of the current study and its implications for practice by extending our analysis to also identify circumstances under which the effect of knowledge about distributed information is more likely to obtain. Testing the effects of a moderator variable that is directly implied by our analysis of the mediating process broadens the empirical basis for our conclusions regarding this mediating process. Moreover, it helps identify measures that could be taken in practice to stimulate groups’ use of distributed information. Our reasoning implies that factors that stimulate group members to use their knowledge about information distribution to inform their task representations should increase the extent to which knowledge about who knows what has positive effects. Based on this, we argue that reflection on the implications of knowledge about distributed information is an important moderator of the effects of knowledge about information distribution.

One of the first to recognize the effect that knowledge of who knows what can have on group performance was Wegner (1987). In his transactive memory theory, he proposed that when a group of people work together for a period of time, they develop a so-called transactive memory system. A transactive memory system refers to a set of individual memory systems in combination with knowledge of who knows what and the communication that takes place between individuals (Wegner, Guilano, & Hertel, 1985). Several studies have shown that transactive memory systems can substantially improve group performance on a variety of tasks (Austin, 2003, Hollingshead, 1998a, Hollingshead, 1998b, Lewis et al., 2007, Liang et al., 1995, Moreland et al., 1998, Moreland and Myaskovsky, 2000). More evidence that knowledge about group members’ roles or expertise can influence group behavior and performance can be found in research on shared cognition about the team. Knowledge about which group member knows or does what within the team has also been referred to as mental models of the team (Cannon-Bowers, Salas, & Converse, 1993). Shared mental models of the team have been found to positively affect coordination (Blickensderfer et al., 2000, Mathieu et al., 2000) and team performance (e.g., Mathieu et al., 2000, Peterson et al., 2000).

When it comes more specifically to decision making in groups with distributed information, knowledge about which group member knows what has also been shown to be beneficial for performance (Littlepage et al., 1997, Stasser et al., 1995, Stasser et al., 2000, Stewart and Stasser, 1995). Stasser et al. (1995) compared information sharing and decision performance in groups that had been informed about which member had what knowledge with groups that were not informed about this. They showed that groups in which group members were informed about members’ areas of expertise exchanged more unique information (i.e., information possessed by only one of the group members) and made higher-quality decisions than groups in which members were not informed. While subsequent studies have consistently replicated the effect of knowledge about the distribution of information (for short distribution knowledge), it remains unclear through which processes and under which conditions it leads to higher elaboration of distributed information (Stasser et al., 2000).

Several possible explanations have been posited for the effects of distribution knowledge, some of which have been shown to account for the effects of distribution knowledge on other types of tasks not involving group decision making with distributed information. The first explanation centers on the idea that distribution knowledge facilitates coordination between team members. Research in transactive memory has shown that members of groups that had developed a transactive memory were better able to anticipate each other’s behavior and were able to work together more smoothly (Liang et al., 1995, Moreland and Myaskovsky, 2000). For decision-making groups with distributed information, this means that when group members know who knows what they can focus on recalling and mentioning information that falls within their own area of expertise. This could then lead to a decrease in the tendency to primarily discuss information that is already known to all before discussion (Stasser, 1999) and an increase in the likelihood that unshared information is mentioned (Stasser et al., 1995). However, Stasser et al. (2000) found that when members had distribution knowledge they did not mention more information that fell within their own area of expertise, although overall more unshared information was mentioned within the group. Thus, although giving group members distribution knowledge led to more information exchange it did not do this by affecting coordination of information sharing (Stasser et al., 2000).

A second explanation posits that information use is affected by an increase in acceptance or absorption of information, because of a greater perceived validity of information that is known to all. When information can be socially verified, it is more likely to be absorbed by other group members (Parks & Cowlin, 1996). Yet, when group members know that an individual who possesses information that cannot be socially verified (i.e., distributed information) is knowledgeable about the topic, the information is more likely to be absorbed, because it is perceived as valid by other group members. The information may then be more likely to be used in making a group decision (Hollingshead, 1998a, Liang et al., 1995, Moreland and Myaskovsky, 2000). Stasser et al. (2000) tested social validation as an explanatory factor by examining retention of unshared information. Although perceived validity has been shown to account for effects of distribution knowledge on an assembly task (Liang et al., 1995) and a recall task (Hollingshead, 1998a), Stasser et al. did not find that people remembered more distributed information when they were given distribution knowledge (Stasser et al., 2000). While we would not argue that absorption may be discarded as an explanatory factor altogether, this shows that at least individual retention of information after group decision making was unable to account for the effect. As we will argue later, it seems feasible that absorption of information indeed plays a role. However, we argue that there is a factor more directly affected by distribution knowledge that leads to more absorption, among others (i.e., task representations).

Thus, while there thus is robust evidence for the effect of distribution knowledge on decision-making performance, it remains unclear which process drives this effect. Addressing this issue, we propose that distribution knowledge elicits task representations that emphasize exchange and integration of informational resources (cf. van Ginkel & van Knippenberg, 2008). Based on this proposition, we also identify a factor that moderates the effect of distribution knowledge because it affects the extent to which distribution knowledge elicits such task representations: reflection on the task.

More or less by definition, decision-making groups have to reach agreement to arrive at a decision. Not surprisingly then, the need to reach agreement and find common ground may assume center-stage in group members’ understanding of the group task. Unfortunately, this often seems to suggest to group members that the appropriate way to make group decisions is to discuss members’ decision preferences and the information that directly supports these preferences rather than more broadly discussing decision-relevant information even when it runs counter to prevailing preferences (van Ginkel & van Knippenberg, 2008). Thus, the focus on the need to reach an agreement may lead group members to overlook information (cf. Hastie and Pennington, 1991, Schulz-Hardt et al., 2006) or to especially consider information possessed by all members rather than distributed information (cf. Stasser & Stewart, 1992; Stewart and Stasser, 1998; Wittenbaum, Stasser, & Merry, 1996). Yet, when group members do have an adequate understanding of the informational requirements of their task, they are likely to rely more on members’ distributed information (van Ginkel & van Knippenberg, 2008).

The concept of task representations was proposed to capture the perception and understanding that group members have of the group decision task. In terms of Newell and Simon’s (1972) theory of human problem solving, task representations can be seen as a part of the problem solving space. Task representations are systems of symbols employed by the human mind that are used for problem solving. Like similar concepts such as mental models they form a simplification of the outside world that guides behavior towards a goal (cf. Cronin and Weingart, 2007, Rouse and Morris, 1986, Tindale et al., 1996). Different task representations may lead individuals and groups to engage with the task in different ways. In addition, some representations may be more task-appropriate than others in that they are more likely to lead to high performance (e.g., Mathieu, Heffner, Goodwin, Cannon-Bowers, & Salas, 2005). van Ginkel and van Knippenberg (2008) proposed that these notions are particularly relevant to our understanding of decision-making groups’ use of distributed information. In an experiment, they showed that when groups were stimulated through experimental instructions to form task representations emphasizing information elaboration, groups discussed more distributed information and made higher-quality decisions than groups that were not stimulated to form such task representations. In terms of Newell and Simon’s (1972) discussion of task representations, such task representations may be seen as consisting especially of operators (rules about how to solve problems) that dictate that the exchange, consideration, and integration of information lead to the goal of high-quality group decisions (cf. Cronin & Weingart, 2007).

We argue that when group members have distribution knowledge, this knowledge may form the basis for task representations that emphasize information elaboration. When group members know that people within their group possess unique informational resources, this knowledge can affect their perceptions of and ideas about their task (cf. Brandon & Hollingshead, 2004). The combination of knowing that all members possess different knowledge and the fact that they are to work on a particular task together as a team, may lead group members to believe that their specific knowledge is valuable for the task at hand. Put differently, the goal of a high-quality group decision in combination with distribution knowledge (i.e., ‘elements’, Newell & Simon, 1972) may lead group members to infer that the right strategy (i.e., ‘operator’, Newell & Simon, 1972) to reach this goal is to exchange and integrate the team’s distributed information. Thus, group members’ distribution knowledge may lead to the adoption of task representations emphasizing elaboration, which in turn should lead to higher group decision-making performance. That is, we argue that it is not coordinated recall of information (cf. Stasser et al., 2000) nor perceived validity of information (cf. Stasser et al., 2000), but a more fundamental change in members’ perception of the task that drives the effect distribution knowledge: Knowledge of who knows what changes group members’ understanding of the task to an understanding that puts the exchange and integration of information center-stage.

van Knippenberg et al., 2004, van Ginkel and van Knippenberg, 2008 defined elaboration as the exchange, discussion, and integration of task-relevant information. Building on this analysis, in the present study we focus on information elaboration rather than on information exchange alone, because several studies have shown that high-quality decision making in groups with distributed information is not solely a result of information exchange (Gigone and Hastie, 1993, Larson et al., 1994, Scholten et al., 2007, Winquist and Larson, 1998), but at least as much a result of what groups subsequently do with the information. Recent studies of decision making in groups with distributed information support the proposition that elaboration is the key process driving decision-making performance (Homan et al., 2007, Kooij-De Bode et al., in press, van Ginkel and van Knippenberg, 2008; cf. Homan, Hollenbeck, Humphrey, van Knippenberg, van Kleef, & Ilgen, in press). Indirect evidence that elaboration may be involved in the effects of distribution knowledge also derives from Stasser et al. (1995) who focused on information repetition, which may be seen as an aspect of elaboration (cf. Scholten et al., 2007). They found that not only the exchange of information but also the repetition of information is positively affected by knowledge about the distribution of information.

Now that we have proposed an answer to the question of which process mediates the effect of distribution knowledge, we can also determine under which circumstances this effect is more likely to obtain. Theory on reflection suggests that the extent to which group members reflect on their team and task may be an important factor in this respect. Within the literature on learning, reflection has been described as a key variable (Boud, Keogh, & Walker, 1985) that involves incorporating an otherwise uncoded experience into a knowledge system by means of engaging in cognitive deliberation (cf. Usher, 1985). Developing task representations based on knowledge about a task may be seen as a process quite similar to learning, which may likewise benefit from higher levels of cognitive deliberation. Consequently, when group members reflect on their task before they start with group decision making, they should be more likely to make use of their distribution knowledge when developing task representations than when they do not reflect and immediately start acting. This also is in line with the literature on reflection at a team level, which suggests that team reflection may help teams develop a better understanding of their task (West, 1996) and more in general can positively affect team process and performance (Carter and West, 1998, Schippers et al., 2003, Tjosvold et al., 2003).

Based on the above, we propose that the process of reflection is instrumental in mobilizing group members’ distribution knowledge. When group members reflect on the team and the task, they are more likely to consider the implications of their distribution knowledge. When they reflect on the existence of distributed information within the group, they may come to realize that exchange and integration of decision-relevant information is key to successful task performance. In other words, they may form task representations that emphasize information elaboration. In contrast, when group members do not engage in such a process of reflection, they are less likely to consider the implications of their distribution knowledge, and are therefore less likely to develop task representations emphasizing elaboration.

Fig. 1 shows a graphic representation of the proposed relationships involved in the effects of distribution knowledge on decision quality. Central to our model is the proposition that distribution knowledge has a positive effect on group decision quality, because this knowledge leads to task representations emphasizing information elaboration. By affecting these task representations distribution knowledge stimulates group information elaboration, which results in higher-quality decisions. Stated as hypotheses, we expect that groups that have knowledge about the distribution of information make higher-quality decisions than groups that do not have this knowledge (Hypothesis 1a). In line with our proposition regarding the mediating process, we further predict that groups that have distribution knowledge are more likely to develop task representations emphasizing information elaboration than groups that do not have distribution knowledge (Hypothesis 1b) and that groups that have distribution knowledge engage in more elaboration of distributed information than groups that do not have distribution knowledge (Hypothesis 1c). In addition, we expect task representations to mediate the relationship between distribution knowledge and decision quality (Hypothesis 2) and we expect that information elaboration mediates the relationship between task representations and decision quality (Hypothesis 3).

In addition, reflection is expected to moderate the effect of distribution knowledge on task representations, because reflection is likely to enable members to form task representations emphasizing information elaboration when they have distribution knowledge. Parallel to what we predicted for the main effects, the moderation effect of reflection is expected to affect group decision quality through its effect on task representations and group information elaboration. Therefore we predict interactions between distribution knowledge and reflection, such that when group members reflect on their task, distribution knowledge leads to higher group decision quality (Hypothesis 4a), is more likely to lead to group members developing task representations emphasizing information elaboration (Hypothesis 4b), and leads to more group elaboration of distributed information (Hypothesis 4c). In addition, the moderation effect of reflection and distribution knowledge on group decision quality is expected to be mediated by task representations emphasizing information elaboration (Hypothesis 5). (Note that the subsequent step in the mediating process—information elaboration—is already implied by Hypothesis 3.)

We put these hypotheses to the test in an experiment in which we manipulated distribution knowledge and reflection on the group task. The experimental nature of the study not only allowed us to reach conclusions about causality, but also allowed us to use audio–video data for a relatively objective and unobtrusive behavioral measurement of elaboration of decision-relevant information (van Ginkel & van Knippenberg, 2008; cf. Weingart, 1997).

Section snippets

Design and participants

The experiment had a 2 (distribution knowledge: yes/no) × 2 (reflection: yes/no) design. A total of 375 students (240 men and 135 women) were assigned to 125 three-person groups. Due to technical problems audio–video data were lacking for six groups. Because there is no reason to suspect that these groups behaved differently from the other groups, the groups were kept in the analyses (listwise deletion was used in the regression analyses involving elaboration). Groups were randomly assigned to

Preliminary analyses

awg(1) Values were calculated for task representations and the knowledge manipulation check to determine the level of analyses (Brown & Hauenstein, 2005). A cut-off score of approximately .70 has been reported as a threshold level (Brown & Hauenstein, 2005). Task representations had a mean awg(1) value of .91 and the knowledge manipulation check had a mean awg(1) of .92, which signal high levels of agreement within groups for both variables. Therefore task representations and the knowledge

Discussion

Decision-making groups with distributed information often make suboptimal use of their informational resources. Knowledge about the distribution of information—knowledge about who knows what—has consistently been shown to have a positive effect on group information exchange and the quality of group decisions, and the present study too confirms this basic finding. This effect presented somewhat of an enigma, however, because it was unclear what caused the effect (Stasser et al., 2000). The

Conclusion

We identified the process responsible for the effects of knowledge about who knows what in decision-making groups with distributed information, and we showed that this analysis provides a point of departure for the identification of factors that may bolster this effect. From this analysis flow a series of potentially important directions for future research, such as regarding the interplay of cognitive representations of the team and representations of the task and the further exploration of

References (72)

  • G.M. Wittenbaum et al.

    Tacit coordination in anticipation of small group task completion

    Journal of Experimental Social Psychology

    (1996)
  • V.Y. Yzerbyt et al.

    Adjusting researchers’ approach to adjustment: On the use of covariates when testing interactions

    Journal of Experimental Social Psychology

    (2004)
  • J.R. Austin

    Transactive memory in organizational groups: The effects of content, consensus, specialization, and accuracy on group performance

    Journal of Applied Psychology

    (2003)
  • R.M. Baron et al.

    The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations

    Journal of Personality and Social Psychology

    (1986)
  • Blickensderfer, E., Cannon-Bowers, J. A., & Salas, E. (2000). The relationship between shared knowledge and team...
  • D.P. Brandon et al.

    Transactive memory systems in organizations: Matching tasks, expertise, and people

    Organization Science

    (2004)
  • R.D. Brown et al.

    Interrater agreement reconsidered: An alternative to the rwg indices

    Organizational Research Methods

    (2005)
  • R.J. Brown et al.

    The utility of experimental research in the study of transformational/charismatic leadership

    Leadership Quarterly

    (1999)
  • J.A. Cannon-Bowers et al.

    Shared mental models in expert team decision making

  • S.M. Carter et al.

    Reflexivity, effectiveness, and mental health in BBC-TV production teams

    Small Group Research

    (1998)
  • M.A. Cronin et al.

    Representational gaps, information processing, and conflict in functionally diverse teams

    Academy of Management Review

    (2007)
  • C.K.W. De Dreu et al.

    Motivated information processing in group judgments and decision making

    Personality and Social Psychology Review

    (2008)
  • R.L. Dipboye

    Laboratory vs. field research in industrial and organizational psychology

  • A. Edmondson

    Psychological safety and learning behavior in work teams

    Administrative Science Quarterly

    (1999)
  • R.J. Ely et al.

    Cultural diversity at work: The effects of diversity perspectives on work group processes and outcomes

    Administrative Science Quarterly

    (2001)
  • D. Gigone et al.

    The common knowledge effect: Information sharing and group judgment

    Journal of Personality and Social Psychology

    (1993)
  • D.E. Gray

    Facilitating management learning: Developing critical reflection through reflective tools

    Management learning

    (2007)
  • A. Gurtner et al.

    Getting groups to develop good strategies: The effects of reflexivity interventions on team process, team performance, and shared mental models

    Organizational Behavior and Human Decision Processes

    (2007)
  • R. Hastie et al.

    Cognitive and social processes in decision making

  • V.B. Hinsz et al.

    The emerging conceptualization of groups as information processers

    Psychological Bulletin

    (1997)
  • A.B. Hollingshead

    Retrieval processes in transactive memory systems

    Journal of Personality and Social Psychology

    (1998)
  • Homan, A. C., Hollenbeck, J. R., Humphrey, S. E., van Knippenberg, D., van Kleef, G. A., & Ilgen, D. R. (in press)....
  • A.C. Homan et al.

    Bridging faultlines by valuing diversity: Diversity beliefs, information elaboration, and performance in diverse work groups

    Journal of Applied Psychology

    (2007)
  • Kooij-De Bode, H. J. M., van Knippenberg, D., & van Ginkel, W. P. (in press). Ethnic diversity and distributed...
  • J.R. Larson et al.

    Discussion of shared and unshared information in decision making groups

    Journal of Personality and Social Psychology

    (1994)
  • Cited by (0)

    View full text