Abstract
Groups rarely use the unique knowledge of their members when making decisions, focusing instead on knowledge that members have in common. This tendency to neglect the expertise of group members severely limits the effectiveness of group decision making. Previously, this problem has been addressed by showing that groups will pool task-relevant information and make effective decisions if members have knowledge of each other’s expertise. However, these studies are generally limited because they disregard why people use each other’s expertise once they are aware of it. The current study uses expectancy theory to investigate this issue and to link motivation to information exchange in groups. Results of a hidden profile study involving 40 groups indicated that expectancy motivation drives groups to use expertise awareness, exchange more unique information, and thus solve a hidden profile problem correctly.
Imagine a group of 19 air traffic controllers regulating a fly zone in which two UH-60 military helicopters are operating. They are part of a mission to provide helicopter transportation for a high-ranking U.N. official from the Iraq–Turkey border to a location deeper into the Iraqi secure zone to meet with Kurdish leaders. In addition, two friendly F-15 fighter planes are on patrol in their fly zone. The team’s job is to monitor the airspace and facilitate communication between the helicopters and fighter planes. However, when the F-15s identify the helicopters as hostile Russian aircraft and call down to air traffic control for permission to fire, those team members responding do not know how to interpret the radar signals. In addition, there is no communication between the radar operators, those members communicating with the helicopters, and those receiving the call from the fighter planes. So, instead of correctly identifying the helicopters as friendly vehicles, they give the planes permission to fire. This is the brief account of how 26 people lost their lives in a friendly fire incident in Iraq on April 14, 1994 (Snook, 2000). Most likely, this disaster could have been prevented if the air traffic crew had been able to correctly identify that some of their members had unique information concerning the helicopters while other members had expertise concerning the tracking apparatus (Snook, 2000). However, even with that knowledge, it would have been necessary to seek out and integrate appropriate information to make the correct informed decision. I argue that the motivation to seek out and connect information has been overlooked in certain paradigms within the group decision-making literature. Unfortunately, this solemn example stands as only one of a myriad of problems that have resulted from inefficient information exchange in groups.
Groups, instead of individuals, are often chosen to make decisions because of their greater knowledge base and the diversity of members’ perspectives. However, a multitude of studies have shown that collectives often fail to exploit their potential because groups are unable to use the specialized knowledge of their members. Instead, group members tend to share information they commonly hold (Mesmer-Magnus & DeChurch, 2009; Wittenbaum, Hollingshead, & Botero, 2004)—a problem that influences efficiency of problem solving. When individuals are aggregated in an attempt to combine knowledge and perspectives, it may result in the discussion of even less information than a single person would use since the common set of information present between two individuals must be less than the amount held by either alone.
Due to the implications of these findings, several avenues of research have begun to examine mechanisms that increase the efficiency of group information exchange (e.g., Galinsky & Kray, 2004; Postmes, Spears, & Cihangir, 2001; Stewart & Stasser, 1995). One of the most fruitful has integrated research on expertise awareness (Lewis, Belliveau, Herndon, & Keller, 2007) with that on group information sharing. Specifically, research has uncovered that when group members are aware of each other’s domains of expertise they are more likely to query about those domains and thus expose and discuss uniquely held information, thereby increasing group problem-solving efficiency (Stasser, Stewart, & Wittenbaum, 1995; Stewart & Stasser, 1995).
Despite the abundance of research supporting the role of expertise awareness in increasing the effectiveness of group decision making (Peltokorpi, 2008), few studies investigate why group members use others’ expertise to the benefit of group decision making. This is a pressing issue since research has found that social factors present in groups can limit group member information exchange, for example, coworkers’ dislike for each other (Casciaro & Lobo, 2005, 2008; Hollingshead, Jacobsohn, & Beck, 2009). In other words, people will not use others’ knowledge if they are not motivated to do so.
Therefore, I propose that in addition to knowledge of who knows what within a group, group members’ motivation to exchange information with each other plays a vital role in group decision making. Expectancy theory states that individuals’ desire for specific rewards (valence), perception of the link between their own actions and receiving those rewards (instrumentality), and belief that they can complete the behaviors necessary to receive those rewards (expectancy) drives their effort toward completing tasks (Vroom, 1964). In groups, members’ instrumentality toward the outcome is attenuated since each group member requires some credit for the group performance. This attenuation results in lower effort exerted toward group tasks (Karau & Williams, 2001). Furthermore, this tendency for decreased instrumentality to reduce group member effort is much stronger in groups that are not task interdependent, or whose members do not believe the entire group is necessary for performance (Wagemen, 1995). Consequently, expectancy theory provides an exemplary vehicle to study the influence of motivation on group information exchange since it has been shown to affect group environments and can be easily tracked through task interdependence.
This issue is investigated in a study that explores both expertise awareness and instrumentality as drivers of group decision making. This is done to provide a mechanism that can explain the development of expertise awareness and to parse out the specific influence it has in group decision making scenarios in light of motivational influences. In addition, this focus on the specific influence of motivation, as compared to expertise awareness, in group decision making will help to broaden the motivated information processing literature. While it is known that the specific motivations that drive individual behavior can influence the way groups process information (De Dreu, 2007; De Dreu, Nijstad, & van Knippenberg, 2008), little work has investigated how the level of group member motivation influences the use of individual information in group decision making.
Therefore, this approach provides a more complete explanation of group decision making because investigating the motivation behind information exchange allows for an explanation of not only how groups make better decisions but why. In addition, using expectancy theory in conjunction with expertise awareness addresses this issue by exploring when individuals seek information and what drives groups to use members’ expertise. Answering these questions is important to providing a more complete picture of group information exchange and will become more important as decision-making groups become increasingly common in organizations (Gully, Incalcaterra, Joshi, & Beaubien, 2002).
Expertise Awareness and Group Information Exchange
As mentioned, groups have a pervasive tendency to neglect the expertise of their members (Mesmer-Magnus & DeChurch, 2009; Stasser & Titus, 1985). However, making group members aware of each other’s expertise attenuates this tendency. For example, three-person groups have been found to solve hidden profile tasks with three possible choices at a rate of about one third, what one would expect due to random chance (Stassser & Stewart, 1992). Hidden profile tasks involve groups making decisions where the correct choice is hidden from each individual member despite the aggregate group having enough total information to choose optimally (Stasser & Titus, 1985). In this way each member receives a subset of information regarding a specific task that is not enough for any individual member to correctly solve alone. Such scenarios provide good environments to study information exchange and decision making because information can be strictly controlled and measured, and exchange is necessary to make effective decisions (Stasser & Titus, 1985).
It has also been found that when the clues are divided such that each group member’s unique information is about one of the three decision outcomes, and the group members are told which outcome every member received unique information about, the solve rate jumps to 61% (Stasser et al., 1995). The common explanation for this finding is that when group members are aware of each other’s expertise they are better able to coordinate the processing of information and thus are better able to make effective decisions since they know where appropriate information lies (Stewart, Stasser, & Wittenbaum, 1995).
While it is recognized that groups need to exchange knowledge to produce tangible outcomes, specific motivational mechanisms that may drive this exchange have been almost ignored. For example, Casciaro and Lobo (2008) provide evidence that interpersonal liking within organizations influences the likelihood of individuals to seek and provide information to each other to the point that individuals will virtually ignore others if their dislike of them is strong enough. Similarly, De Dreu and colleagues (2007, 2008) show that groups who are more motivated to understand their surroundings exchange more information; however, this is done in the absence of expertise awareness. Moreover, evidence exists that egocentric goals and distrust of others can limit information exchange within organizations despite the existence of expertise awareness (Javernpaa & Majchrzak, 2008). These findings indicate that some variance in the impact of expertise awareness on group decision-making efficiency may be explained by group members’ motivation to use those systems. Expectancy theory provides a platform to examine this motivation and thus explain instances when individuals will be likely to use expertise because they desire to do so.
Task Interdependence, Instrumentality, and Information Exchange
Being in a group can affect people’s instrumentality in a few ways. Either the presence of other group members can limit one’s own sense of instrumentality since others who help to complete a task require some credit for performance, or, alternatively, other group members can exacerbate one’s own sense of instrumentality by enabling a focal individual to perform. Both reactions can be explained by expectancy theory although each has distinctive implications for group information exchange. In addition, I hypothesize that the way other group members affect one’s own instrumentality depends on the task interdependence of his or her group.
Groups can be more or less task interdependent based on how necessary each member is perceived to be for the achievement of the group outcome (Wagemen, 1995). When groups have low task interdependence, members perceive each other to be unnecessary for performance, eliciting a situation where a few skilled team members can be responsible for the performance of the entire team. In such situations group members impede each others’ instrumentality toward the group outcome. This occurs because when one member completes a subtask and thereby takes responsibility for it, no one else can complete that particular subtask; thus other group members cannot be instrumental toward the group outcome through that particular means (Karau & Williams, 2001). So, in undertakings that are task noninterdependent when one member completes one of the nonspecialized task-related behaviors responsible for performance, no one else needs to. In such situations it is not necessary for team members to share information with each other because each member impedes every other’s instrumentality toward the group outcome (Karau & Williams, 1993). In fact, evidence exists that when tasks are perceived as having low task interdependence individuals are likely to develop their own knowledge structures without taking other group members into consideration (Wagemen, 1995). In addition, group members in task noninterdependent groups are less likely to seek help from each other (Cleavenger, Gardener, & Mhatre, 2007), and are less likely to exchange task relevant information (Katz-Navon & Erez, 2005).
Conversely, in task interdependent groups each member needs each other member to be instrumental to the group outcome, since each member believes that each other member is necessary for group performance (Wagemen, 1995). In other words, in tasks high in task interdependence each member is necessary to each other member’s instrumentality since having one incompetent member makes one’s own effort moot. So, because individuals desire to exert effort in tasks with worthwhile rewards (Bandura, 1997), in task interdependent endeavors it becomes necessary for members to provide information to, and seek information from, other group members since all are vital to team performance. Indeed, the same studies previously cited have also found that members of task interdependent groups seek help from each other (Cleavenger et al., 2007) and exchange task-relevant information with each other (Katz-Navon & Erez, 2005).
These findings provide evidence of a link between instrumentality and information exchange in groups; however, they do not explore the type of information being exchanged. Specifically, they do not investigate whether the information being exchanged is relevant to a focal problem and thereby do not investigate the benefits of aggregation in decision making. In the hidden profile task being used, this problem-relevant information takes the form of the unique information each member holds concerning the task. So, because task interdependence has been shown to lead to the consideration of characteristics of individual group members (Cleavenger et al., 2007; Katz-Navon & Erez, 2005) and information exchange between group members, it is expected it will lead to the exploration of group member expertise. Therefore, the following hypotheses are made:
Hypothesis 1 (H1): Task interdependent groups will exchange more unique information than task noninterdependent groups
Hypothesis 2 (H2): Task interdependent groups will make more optimal decisions than task noninterdependent groups
Method
Participants
One-hundred and twenty students composing 40 three-member groups participated in the study. These students were recruited from various organizational behavior and psychology classes at a large northeastern university and participated in the study for extra credit (63 male, 57 female, Mage = 19.89 years).
Procedure
Participants were divided into four conditions based on the manipulation of task interdependence and expertise awareness. All groups completed the same murder mystery hidden profile task. To do this, participants first individually read a 27-page booklet describing the circumstances of a murder. The materials included a map, a handwritten note, a newspaper article, and numerous interviews. Among these materials were 24 clues containing either incriminating or exonerating evidence toward each of three subjects labeled E, B, and M. Overall, there were six clues incriminating each subject and three clues each exonerating B and M. Specifically, the three clues incriminating E and the six clues exonerating B and M were critical to solving the murder. These nine critical clues were the ones that created the hidden profile and were distributed three each to each team member in a manner such that each team member received the three clues critical to one of the suspects. Participants were then aggregated into their three person groups and asked to solve the mystery. Groups were told they could take as long as they wanted to discuss the case, but that they must come to consensus regarding the killer and that all groups to correctly solve the case would be entered in a raffle to win US$20 per group member. All discussion was videotaped.
Finally, after submitting their answer, each group member completed a questionnaire composed of an expertise awareness manipulation check and a task interdependence manipulation check. The expertise awareness check asked each member how much information each group member had, compared to the other members, about each suspect (1 = much less than other members, 4 = the same as other members, 7 = much more than other members). Groups who correctly identified the specific suspect all group members had the most information about were coded as having expertise awareness and those that could not were not. The task interdependence manipulation check asked participants how much the outcome was dependent on themselves (0%-100%) and on each of their teammates (0%-100%). Finally, two steps were taken to ensure the results found were due to this instrumentality difference and not valence or expectancy, which were the two components of expectancy motivation not hypothesized to be responsible for the results found. First, valence, the desirability of a reward linked to a specific outcome (Vroom, 1964), in this case the reward for correctly solving the mystery, was held constant. As stated, in each condition group that correctly solved the mystery were entered into a raffle for US$60. In addition, expectancy was measured and compared between conditions. Expectancy refers to the belief that one’s effort will result in a desired performance (Vroom, 1964). Therefore, as an expectancy measure, individuals were asked how confident they felt in their ability to successfully solve the mystery on a scale of 0% to 100%.
Manipulations
As mentioned, the first dimension manipulated was task interdependence. To enact this manipulation a procedure was borrowed from Stasser et al. (1995). Specifically, participants in the high task interdependence condition were told, in the presence of other group members, that each one of them had received unique information regarding the case and that each group member would be necessary to solving the case. There were two distinct conditions with high task interdependence. In the first, group members were told which suspect they had the most information about in front of each other group member (Expert Public). This was intended to create expertise awareness within the group. In this condition, group members were each told, in front of each other,
Your group is about to complete a murder mystery task. Please use the information given to attempt to deduce which suspect has perpetrated the crime. In addition, each of you will be given an expert role about a specific defendant and be given information about that defendant that no one else has.
In the second task interdependent condition, participants were given the roles of lead detective, forensic specialist, and eye witness, in front of each other. These roles were chosen because they pertain to the environment surrounding a murder investigation. However, other than this purpose, they were completely random (Random Roles). In other words, they had nothing to do with the information contained in each member’s packet. In this condition group members were told together,
Your group is about to complete a murder mystery task. Please use the information given to attempt to deduce which suspect has perpetrated the crime. In addition, each of you will be given a role for the case and will be given information in accordance with that role that no one else has.
This second task interdependent condition was novel and was created for the current study. Therefore, two high task interdependent conditions were created, one whose groups knew whom each member received more information about, and one who only knew they needed each member to complete the task.
In addition, and again following the work of Stasser et al. (1995), one task noninterdependent group was created with expert roles. In this condition, members were told individually that they were experts on the suspect for whom their packet contained more information (Expert Private). Therefore, they were told the same thing as the expert public group, only they were told individually. Furthermore, a control condition was created with no roles and where participants were not told all members were necessary for performance (Control).
Coding
First, the group discussions were transcribed from the videotapes that were recorded. Three raters read the 40 transcripts and classified each based on, among other things, the amount of unique information exchanged (out of the nine total pieces; interrater reliability [IRR] = .92), the number of times group members disagreed, and the number of times group members made information requests of each other. Coders were unaware of any hypotheses at the time they coded the narratives. Specifically, disagreements were defined as any time a teammate specifically told another teammate their idea was incorrect (IRR = .91). For example, in one group, group member A stated, “I think we can rule out (suspect A), he’s just too stupid.” Then group member B responded, “I don’t know about that, he was the last one to touch the [murder weapon].” This clear retort was coded as a disagreement. In addition, information seeking behaviors were defined as any time a teammate asked another teammate for a specific piece of information relevant to the case (IRR = .94). For example, one participant asked his group, “What did your packets say about [suspect A]?” Disagreements in classification were resolved by discussion between the coders.
Results
Manipulation Checks
Two manipulation checks were conducted. To do these, hierarchical linear modeling was used because a group-level manipulation was enacted to account for variance in an individual-level outcome. This setup violates the assumption of independence of observation, made in ANOVAs; however, hierarchical linear modeling (HLM) techniques provide unbiased estimates of standard errors for hypothesis testing in these situations (Raudenbush & Bryk, 2002; Snijders & Bosker, 1999). First, a hierarchical null model (i.e., a model with no predictors at either level) was created to examine how much variance in the dependent variable could be predicted by participants’ perceptions of task interdependence. The intraclass correlation (ICC), which measures the level of interdependence of the data within nesting unit, was .72, indicating that 72% of the variance within task interdependence can be explained by within-groups differences. Then, the participants’ condition was added to the model to assess whether the task interdependence manipulation was successful. As expected, those in the two task interdependent conditions (M = 78.92, SD = 44.47) believed choosing a suspect was more dependent on their teammates than those in the control and expert private conditions (M = 59.57, SD = 38.05), even when controlling for group membership, βTaskInterdependence = 19.35, t(38) = 2.57, p < .05. Second, expectancy was compared between all conditions. As expected, HLM indicated that condition did not significantly affect team members’ belief that they, individually, had the ability to solve the murder mystery, βCondition = −0.51, t(38) = −.41, p = .68, ICC = .78. Therefore, the results found cannot be linked to differences in the base ability beliefs of individual team members.
Because the hypotheses involved group-level dependent variables, the amount of unique information exchanged by a group as a whole and whether a group solved the problem correctly, HLM could not be used to test them. Instead, ANOVAs were used. Hypothesis 1 predicted that those in the task interdependent conditions would exchange more unique information regarding the case. An ANOVA with a post hoc Tukey test indicated that this was the case. First, an ANOVA indicated that differences in the amount of unique information shared existed between conditions, F(3, 36) = 3.33, p < .05. A subsequent Tukey test then indicated that the amount of unique information shared in both the public expert roles (M = 6.27, SD = 1.62) and random roles conditions (M = 5.94 SD = 1.96) was greater than the amount of information shared in the private expert roles (M = 4.11, SD = 1.73) and control conditions (M = 3.45, SD = 2.61), see Table 1. This finding, along with not finding a difference between the two task interdependent conditions indicates that motivation does play a significant role in information exchange in these sorts of problems and that expertise awareness cannot tell the entire story of within-group unique information exchange.
Mean Differences in the Murder Mystery Hidden Profile
Note: Those cells in all rows except for solve rate with different superscripts have significantly different means, p < .05. Those cells in the solve rate row with different superscripts have marginally different means, p < .06.
In addition, it was hypothesized that the same two task interdependent conditions would solve the mystery more often than the conditions with low task interdependence. Seven of 10 groups in both the task interdependent conditions solved the mystery correctly, for a total of 14 out of 20, as compared to 4 of 10 groups in both the task noninterdependent conditions, for a total of 8 out of 20. A chi-square analysis confirmed these solve rates were marginally different, χ2 = 3.64, p < .06, partially supporting Hypothesis 2, see Table 1. Like the previous finding, the clear difference between task interdependent and noninterdependent groups coupled with the nonexistence of any difference in solve rate between the task interdependence groups indicates that motivation plays a key role in group decision making and that expectancy theory can help to explain this role.
Then, the role of task interdependence as a driver of these findings was investigated. First, the aggregate level of instrumentality that each group member felt towards the task did not differ between those in task interdependent and noninterdependent groups, t(38) = .47, p = .50. Combined with the task interdependence manipulation check, this finding supports the theory that those in task interdependent groups allow themselves to feel their teammates are instrumental to group outcomes, without attenuating their own instrumentality. Hierarchical linear modeling further supported this premise. In the two task noninterdependent conditions, participants’ own instrumentality inversely predicted how instrumental they believed their teammates were to the group task, βInstrumentality = −0.50, t(54) = −3.43, p < .01, ICC = .78. However, no such relationship existed in task interdependent groups, βInstrumentality = 0.16, t(54) = 0.69, p = .50, ICC = .78. Therefore, it appears that in task noninterdependent groups, because group members view each other as impediments to their own instrumentality, they cannot allow themselves to believe that both they and their teammates are highly instrumental to group performance. However, task interdependent group members do not suffer from this issue. This finding is important because participant’s own instrumentality was found to positively predict unique information sharing across conditions, βInstrumentality = 0.25, t(106) = .264, p =.01, ICC = .71.
Additional Analysis
In order to further verify the influence of expectancy motivation on expertise awareness development, analysis of differences in the number of groups which had expertise awareness, between conditions, was conducted. As stated, having expertise awareness was defined as a group in which each member correctly identified which suspect they, and each other group member, received the most information about. First, the greatest number of expertise aware groups, 70%, were in the public expert condition. This could be expected as these groups were instructed which member had what information about which suspect. The second greatest number of expertise aware groups, 50%, were formed in the random roles condition. In line with the logic behind Hypotheses 1 and 2, this is believed to be because people in the random roles condition were motivated to exchange information with each other, allowing their expertise to be made salient to the group. Despite each member of the private expertise condition knowing one third of who knew what, by being told that he or she was an expert on one suspect, only 40% of the groups in that condition had expertise awareness. Finally, only one group in the control condition had expertise awareness. Chi-square analysis indicated that the pattern of these differences was significant, χ2 = 7.67, p = .05, see Table 1. Still, it is important to note that, although as many private expert groups had expertise awareness as random roles groups, random roles groups still solved the problem at a significantly higher rate. In combination this evidence lends further support to the proposition that merely having expertise awareness is not enough; group members must be motivated to use it. Task interdependence provided such motivation in the current study.
Since the random roles condition facilitated expertise awareness and produced effective decisions, the investigation of the mechanisms that may have played a role in these relationships warrants investigation. Previous literature has tended to neglect this process information in favor of solely focusing on the unique and common information exchanged by groups (Lu, Yuan, & McLeod, 2011). To this end, differences in other information exchange behaviors were analyzed in reference to the conditions created. Specifically, disagreements with teammates and information requests were investigated as these two behaviors were assumed to be integral to converging on an appropriate conclusion.
Differences between conditions appeared in both behaviors (Disagreement, F(3, 36) = 7.82, p < .001; Information Seeking, F(3, 36) = 41.79, p < .001) with Tukey tests indicating those groups in the random roles condition averaging more disagreements (M = 4.22, SD = 2.59) and information requests (M = 3.89, SD = 2.86) than those in the other conditions. Concerning disagreements, those in the public expertise condition disagreed with each other more (M = 2.10, SD = 2.18) than those in the private expertise (M = 1.01, SD = 1.16) and control conditions (M = .70, SD = .78), see Table 1. The amount of information requests in the expert public (M = 2.03, SD = 2.04), expert private (M = 2.00, SD = 1.69), and control conditions (M = 2.30, SD = 2.72) were equivalent, see Table 1.
Again, these results point to specific motivated behaviors that may be responsible for the development of expertise awareness, as both task conflict (Jehn, 1995) and the seeking and sharing of information (Mesmer-Magnus & DeChurch, 2009) have been found to greatly increase the performance of decision-making groups, especially when making complex decisions. Based on this research it is surprising that groups in the public expert condition both disagreed with each other less, and made fewer information requests, than those in the random role condition, considering both had high task interdependence. However, it may be that once groups have expertise awareness group members do not need to disagree as much as they concede points to respective experts. Still, those in the public expert condition disagreed with each other more than those in the two task noninterdependent conditions. In addition, once individuals are aware they are experts in a specific area they may volunteer applicable information, making information requests less necessary.
Finally, the exchange of common information was investigated. Because task interdependence was used to increase group member instrumentality, and thus motivation, it is possible that groups made better decisions because they shared more total information, instead of solely unique information. To test this possibility, an ANOVA was run and indicated that differences in the amount of common information shared did exist between conditions, F(3, 36) = 4.95, p < .05. Specifically, those in the two outcome interdependent conditions shared more common information (M = 32.61, SD = 19.21) than those in the two other conditions (M = 20.48, SD = 13.27). However, a subsequent logistic regression indicated that only the sharing of unique information led to increased group problem solving, β = .314, SE = .129, p < .05. Furthermore, unique information sharing predicted 31% of the variance in whether groups solved the hidden profile correctly, whereas the amount of common information shared was not a significant predictor of group problem-solving efficiency, β = −.059, SE = .035, p = .094. Therefore, although outcome interdependence increased all types of information sharing, it was the increase in unique information sharing that allowed groups to make more efficient decisions. In addition, a post hoc Tukey test indicated that those in the random roles condition (M = 36.56, SD = 23.59) exchanged more unique information than those in the expert public condition (M = 28.67, SD = 12.52). This indicates that, although groups with higher motivation have the tendency to exchange more common information, this tendency can be attenuated by expertise awareness.
Discussion
The current study was conducted to build on and extend previous findings concerning information exchange and expertise in hidden profile groups. Specifically, the role of expectancy theory and individual motivation in the development and use of expertise awareness was investigated to explore why expertise is used and to integrate these lines of research. To do this a hidden profile task was used and both task interdependence and expertise awareness were manipulated.
Initially, I found that only the task interdependence manipulation, which is indicative of perceived instrumentality within groups (Karau & Williams, 2001; Wagemen, 1995), influenced the mention of unique information, supporting Hypothesis 1. This finding clearly points to shared instrumentality as a driver of information exchange in groups, and is important for a few reasons. First, it is important because it shows the motivational premises described in expectancy theory can have significant effects on not only how much effort individuals put into group discussion but also what they do with that effort, in terms of the type of information they share. Specifically, increasing instrumentality through task interdependence increased the sharing of both common and unique information, the latter of which led to improved group decision making. Second, this finding set the stage for investigating Hypothesis 2, which deals with the motivation to use expertise.
The data supporting Hypothesis 2 indicated that task interdependence was a necessary condition to increase the solve rate in the current study. This finding replicated previous research describing the positive influence of seeing each group member as important to the group outcome on information seeking and giving (Cleavenger et al., 2007; Katz-Navon & Erez, 2005). It also provides a fuller understanding of the importance of assigning experts within groups to facilitate information exchange. Specifically, assigning experts is important to both engender the motivation to exchange full information and to create awareness of others’ expertise. In addition, the findings indicate that even in the absence of experts, assigning nonoverlapping roles to group members in order to create a nonzero sum perception of instrumentality drives group members to seek other expertise relative to the task.
Furthermore, a posttest indicated that those groups high in task interdependence who were given random roles were still aware of each other’s expertise half of the time, as compared to 70% of the time when groups were told who know what as a group, and 40% of the time when group members were individually told their expertise. This finding provides evidence that task interdependence is enough to promote the creation of expertise awareness, even in the absence of preexisting information regarding expertise. Finally, part of the reason for this was that those in random roles groups were more willing to disagree with each other and request information from each other than in the other conditions. In light of the expectancy theory hypotheses, one reason for the difference between the random roles and expert group conditions in these two behaviors could have been that expert group members had no need to disagree as they knew who was an expert on which suspect. In addition, requesting information may have been less necessary since members may have readily volunteered information on their expertise.
Limitations and Directions for Future Research
As is the case with any study, certain design characteristics were limited in favor of others. First, the hidden profile case is a fabricated scenario where complete information regarding the problem is present. This setup provides an excellent task to study group decision making since all relevant information can be accounted for. Still, it would be interesting to study the influence of instrumentality perceptions with field groups as it is likely all information regarding a problem will not be available to them. In such scenarios the role of expectancy motivation may be even greater since members must seek information which is not immediately available. In addition, the experiments described took place over an hour time period in the lab. These time limitations of the lab precluded the current study from incorporating much information on team development. Yet many processes that occur over time, during interactions, are important to team information exchange and decision making. Finally, a task interdependence intervention was used to influence the relative perceived instrumentality of a focal individual and his or her teammates. Nevertheless, this substitution may not be realistic in all cases. All group members entered the experimental task environment and were told whether the task would be outcome interdependent. This setup allowed the task interdependence manipulation to influence perceptions of teammates’ instrumentality while allowing focal participants’ instrumentality to remain high. However, in some real world groups, task interdependence is not so well defined. Therefore, it may be possible for task interdependence and instrumentality to act independently. For example, mystery solving teammates of Sherlock Holmes, with no information concerning task interdependence, would rate him as being highly instrumental to the group outcome, despite rating the task as low on task interdependence. In such situations it is unlikely the team would share much unique information because they would simply rely on Sherlock to solve the case. Therefore, the results should be taken out of contexts where task interdependence is well defined.
Still, the results show that task interdependence is important in group decision making, and should be applicable in environments that require the constant searching out of task-relevant information, including those striving for creativity and those involving adaptation. Creativity is generally defined as the recombination of associative elements into new ideas that are both novel and useful (Amabile, 1988, 1996; Oldham & Cummings, 1996). In groups, this process involves maximizing member expertise and unique knowledge so members may innovatively apply their diverse experiences to problems (Baruah & Paulus, 2009). The results of the current study indicate that this process can be enhanced by focusing attention on the motivation to search out such information, specifically by assigning roles to team members. Therefore, it may behoove those attempting to use groups to find creative solutions to problems to assign such roles, or, at least, assure every group member feels instrumental in the group’s performance.
Furthermore, the consistently changing business environment has made workplace learning and continuous improvement a modern necessity (Salas & Cannon-Bowers, 2001). This means that a group having the metaknowledge to know how to adjust to new tasks and environments is of great competitive advantage. A main component of adjusting to changing task environments involves seeking applicable information concerning those environments (Bell & Kozlowski, 2010). The results of the current study indicate that this process will be augmented by focusing on the instrumentality perceptions of group members, and making sure each member feels responsible for the group outcome, so they are motivated to search out such task relevant information. Groups improperly motivated may, instead of searching out task-specific information, use similar problem-solving strategies to complete dissimilar tasks, resulting in suboptimal decisions.
Finally, evidence exists that individual’s personal motives influence their information exchange in group environments and thus group decisions (De Dreu et al., 2008; Hollingshead et al., 2009; Steinel, Utz, & Koing, 2010). For example, group members competitively striving toward personal rewards share low amounts of unique information, resulting in poor decision quality (Toma & Butera, 2009). Likewise, the desire to obtain status or to gain credit for group performance can influence information exchange (De Dreu et al., 2008). The current study provides evidence that assigning roles, and thus increasing the task interdependence of group members, may help to attenuate such strategic information sharing since increasing task interdependence creates a situation in which members’ personal goals are dependent on other team members (Karau & Williams, 2001).
Conclusion
It is hoped the current study is seen as a step in integrating the motivation and information exchange literatures. Group decision making is wrought with errors, some of which, as with the friendly fire disaster cited in the introduction, can have dire consequences. In addition, more common, yet just as inefficient, organizational decisions can lead to lower performance and profit. It is clearly important to create knowledge networks within groups to optimize their decision-making ability; however, it must be recognized that individuals must be driven to use these networks and to solve problems in groups, just as they must be individually. Expectancy theory provides one such explanatory mechanism and its investigation, along with other motivational mechanisms present in groups, will help to provide a fuller understanding of group information exchange.
Footnotes
Research Presented at the 2011 Conference of the Interdisciplinary Network for Group Research
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author received no financial support for the research, authorship, and/or publication of this article.