Abstract
Organizations increasingly use teams to perform tasks that are too complex for individuals to perform by themselves (see, for example, Mathieu, Maynard, Rapp, & Gilson, 2008). Teams are small groups of individuals who share a responsibility for specific outcomes (Hollenbeck, Beersma, & Schouten, 2012; Ilgen, Hollenbeck, Johnson, & Jundt, 2005). To perform effectively, team members need to share and integrate information to generate a shared understanding of both their task and their environment (De Dreu, Nijstad, & van Knippenberg, 2008; Mathieu et al., 2008; van Ginkel & van Knippenberg, 2008). However, beyond purely task-related facts and input from the team’s environment, team members are confronted with another source of information, that is, information about the affective states of their fellow team members (Heerdink, van Kleef, Homan, & Fischer, 2013; Van Kleef, 2009; Van Kleef, De Dreu, & Manstead, 2010). For example, while interactively performing a team task, positive feelings, like excitement and happiness, might emerge. Alternatively, during group interactions, negative feelings, such as anger and anxiety, might be expressed. Moreover, several studies have shown that group members converge in their affective experiences during group interaction (see Barsade & Knight, 2015, for an overview; George, 1990; Totterdell, Kellett, Teuchmann, & Briner, 1998).
How should team members respond to the affect expressed in their team to optimize team performance? It is generally believed that, just as having a correct and shared understanding of their task and environment (Mathieu, Goodwin, Heffner, Salas, & Cannon-Bowers, 2000; van Ginkel & van Knippenberg, 2008), team members should also have a sense of how their teammates are feeling. This is because team members need to coordinate their efforts, and adequate behavioral reactions require that team members perceive and understand others’ current states. Doing so is affected by their ability to take another person’s perspective, or the capacity to entertain the point of view of others (Davis, 1980, 1983). For instance, understanding other’s affective states enables people to respond adequately to social situations (Bechtoldt, Rohrmann, De Pater, & Beersma, 2011). Moreover, taking others’ perspectives has been shown to increase negotiators’ abilities to construct win–win agreements (Galinsky, Maddux, Gilin, & White, 2008; Trotschel, Huffmeier, Loschelder, Schwartz, & Gollwitzer, 2011) and facilitates creative idea generation in teams (Hoever, van Knippenberg, van Ginkel, & Barkema, 2012). In other words, perspective taking helps individuals to predict the behaviors and reactions of others, and facilitates smoother and more rewarding interpersonal relationships (Davis, 1983). Therefore, perspective taking is generally viewed as an important asset for interpersonal relations (Galinsky & Moskowitz, 2000; Hoever et al., 2012).
However, we question whether it is always functional for team members to engage in perspective taking when engaged in a team task. Perspective taking leads team members to process information on the affective states of team members as if it were their own, and we argue that when teams’ negative affective states are high, perspective taking deteriorates team performance. Below, we will develop this hypothesis and present the results of a study that tested it.
Group State Affect
Affective states are defined by Russell and Barrett (1999) as “the most elementary consciously accessible affective feelings” (p. 806, see also Forgas & George, 2001). Affective states are more diffuse than emotions and most researchers that have studied affective states distinguish between positive and negative affect (Barsade & Gibson, 2007). Several studies have shown that in interacting groups, affective states tend to converge due to contagion effects and exposure to similar events during interaction (Barsade, 2002; George, 1990; Totterdell et al., 1998). According to the social functional perspective of emotions (Keltner & Haidt, 1999), such shared affect is functional for groups. Specifically, this perspective asserts that sharing each other’s feelings in the context of a group may have become important in human evolution because it serves as a starting point for social bonding and coordinating task performance. As such, social integration, defined as the relational bonds that link team members to each other and to their group (Knight & Eisenkraft, 2015), has been considered a key mediating mechanism that explains the relationship between team affective state and team performance.
However, affective states can have a positive or a negative valence, and although the positive relationship between positively valenced group affect and group performance seems to have been well-established in empirical studies (see, for example, Barsade & O’Neill, 2014; Chi, Chung, & Tsai, 2011; Grawitch, Munz, & Kramer, 2003), there is much less clarity about the relationship between group negative affect and performance (Menges & Kilduff, 2015). Specifically, a meta-analysis (Knight & Eisenkraft, 2015) demonstrated that shared positive affect is positively related to team performance because it benefits groups’ social fabric. Shared positive affect relates to trust, prosocial behavior, commitment, and reduced conflict, which subsequently facilitates performance on team tasks. In contrast, results regarding negative affect were much less clear-cut, which led Knight and Eisenkraft (2015) to the conclusion that the effects of group negative affect on social integration and performance are context-dependent. Shared negative feelings caused by a source external to the team were found to strengthen social integration and performance, but shared negative feelings that emerged from within the group were found to weaken social integration and performance. Also, for groups engaged in a task over time, shared negative affect was negatively related to performance, and the authors suggest this is because shared negative affect led group members to appraise the group itself as something from which to withdraw.
Based on these findings, Knight and Eisenkraft (2015) argue that responses to group negative affect can be seen as analogous to individuals’ responses to stress. A short-duration stress response, especially when derived from an external stimulus, can help mobilize resources and therefore be functional to performance. For example, a stress response can help an individual to run away when being attacked by a lion, or, in a work context, to react quickly to an urgent customer question. However, the stress response is not helpful when it is caused by a factor internal to the individual (e.g., low emotional stability) and when it lasts for a longer time (e.g., when a person faces a work overload for an extended time). Under these circumstances, stress is not conducive to performing the tasks at hand and may negatively influence the individual’s performance (see, for example, Pruessner, Hellhammer, & Kirschbaum, 1999, see also Lazarus & Folkman, 1984). Similarly, negative affect in groups may be useful to mobilize group resources when it is attributed to short-duration, external stimuli—provided that it does not absorb cognitive resources that cannot be expended elsewhere.
Perspective Taking
The above overview shows that the effects of negative affect in groups depend on boundary conditions. Team perspective taking might constitute such a boundary condition. Perspective taking is the process of imagining the world from another’s perspective and seeing things from this other person’s point of view (Davis, 1983). Perspective taking is an inherently cognitive process whereby a person actively thinks about the cognitions, motivations, or emotions of another (Caruso, Epley, & Bazerman, 2006). Perspective taking differs from empathy, as the latter refers to an other-focused affective response that allows someone to emotionally connect with another person (Galinsky, Maddux, Gilin, & White, 2008; Ku, Wang, & Galinsky, 2015; Wang, Tai, Ku, & Galinsky, 2014). It has been argued that when someone takes the perspective of another, he or she perceives the other as more similar to the self (Cialdini, Brown, Lewis, Luce, & Neuberg, 1997). Perspective taking is thought to have a generally positive impact on interpersonal relationships through this perceived similarity between the self and the other (Galinsky, Ku, & Wang, 2005). Thus, when taking another person’s perspective, the focal person thinks about the other’s perspective as if it were his or her own.
As briefly described earlier, there are several studies that speak to the positive effects of perspective taking. Much research in this area concerns the effect of perspective taking on individual behavior (e.g., Axtell, Parker, Holman, & Totterdell, 2007; Galinsky & Moskowitz, 2000; Grant & Berry, 2011). However, perspective taking has also been demonstrated to have positive effects at the dyadic and team levels of analysis. For instance, prior research has found that perspective taking positively influences success in negotiations because it increases negotiators’ ability to discover hidden agendas (Galinsky, Maddux et al., 2008; Trotschel et al., 2011). This was found both when people’s habitual tendency for perspective taking was measured and when it was manipulated. Perspective taking allowed negotiators to create more win–win agreements with better collective outcomes as well as better individual outcomes. Moreover, Hoever and colleagues (2012) demonstrated that instructing research participants to take the perspective of other team members in informationally diverse teams led these teams to be more creative. As these examples show, taking others’ perspective can lead to positive group-level outcomes.
The consequences of perspective taking have been examined in studies that both experimentally manipulated perspective taking as well as in studies that measured it. A number of articles also report studies that include both manipulations and measures (e.g., Galinsky, Wang, & Ku, 2008; Ku, Wang, & Galinsky, 2010). According to Ku and colleagues (2015), who reviewed research on perspective taking, studies that have measured and manipulated perspective taking have yielded similar results. In the current article, we are interested in the team-level effects of habitual perspective taking of team members as a potential moderator of the team affect to team performance relationship. As habitual perspective taking is an individual difference that varies across team members, we expect variation (rather than consensus) within teams regarding perspective taking when team members are not selected on this specific variable. Therefore, to capture the effects of perspective taking at the team level, aggregation cannot be based on a direct consensus model (Chan, 1998). In this case, scholars have argued that the appropriate aggregation of individual difference variables depends on the task context within which a team functions (Barrick, Stewart, Neubert, & Mount, 1998; LePine, Hollenbeck, Ilgen, & Hedlund, 1997; Moynihan & Peterson, 2001). As this perspective is rooted in the theoretical work of Steiner (1972), we used his taxonomy (i.e., disjunctive, conjunctive, and additive tasks) as the basis for our aggregation procedure for perspective taking to the team level (see the section “Method” for more information).
In many team tasks in organizations, the rationale for having teams rather than individuals perform the task is that by being responsible for team outcomes together, all team members can contribute to the team’s success. Many organizational tasks are thus structured to be additive. Hence, we examined an additive task in the current study. As such, we examine the effects of team-level perspective taking on the basis of an additive theoretical model (Chan, 1998), arguing that the extent to which members within a team on average score higher or lower on habitual perspective taking matters for team performance.
An important question, however, is whether team-level perspective taking constitutes a resource or, in contrast, a burden for teams. As discussed above, taking the perspective of others can be positive for group-level outcomes. This is especially the case when others’ affective states contain crucial information that fosters making effective decisions. An example is a negotiation task, where affect expressed by counterparts conveys information regarding whether negotiating dyads are moving toward agreement, or rather, to a stalemate or even conflict escalation. If a counterpart is expressing anger, or disappointment, this conveys important information about the negotiation process and possible outcomes (Van Kleef et al., 2010). As such, research has shown that in negotiations, win–win solutions are facilitated by taking the counterparts’ perspective (Galinsky, Maddux, et al., 2008).
However, perspective taking is also a cognitively demanding activity, because it relies more on controlled processing as opposed to automatic processing (Rossnagel, 2000). As such, it could absorb cognitive resources that cannot be expended elsewhere. Thus, in cognitively demanding tasks, team members face a choice of where they expend the majority of their attention and cognitive resources. In tasks where the affective states of team members convey crucial information that is conducive to effective performance in the task, perspective taking may be worth the effort, and help teams to perform well. When negotiating a joint team agreement, for example, thorough understanding of why team members are angry or disappointed about certain alternatives can help teams craft more effective decisions (Galinsky, Maddux, et al., 2008; Van Kleef et al., 2009). In tasks in which it is vital for team members to focus their attention on their task environment rather than on their team members, taking the perspectives of other team members may become a liability—particularly when others’ affective states do not convey critical information to complete the task at hand. A medical team coordinating a complicated surgery with unexpected problems is arguably not served by team members exploring the deeper causes of its members’ shared affective states, but rather by an undistracted focus on completing the surgery. Likewise, a military team engaged in a strike or attack is likely to be better off if team members focus on coordinating their actions without devoting much energy to understanding the causes of their fellow team members’ affective states. In these cases, taking the perspective of other team members may not be functional, because team members pick up distracting information. However, team members may not be able to effectively adjust this behavior when they have a strong habitual inclination to take others’ perspective.
Following Knight and Eisenkraft’s (2015) stress analogy discussed above, taking team member’s perspective may be especially distracting in groups characterized by negative affective states. As Knight and Eisenkraft’s meta-analysis demonstrated, negative affect in groups is only helpful when it is attributed to factors external to the team and when the situation causing it can be quickly resolved. Teams composed of members high in perspective taking, however, will process group members’ affect as if it were their own (Caruso et al., 2006; Galinsky et al., 2005). Rather than focus their effort and energy on the task at hand, they devote their energy to processing their team members’ negative affective states, thereby causing the exact adverse effects that negative affect in groups can have on performance if team members’ cognitive resources are needed elsewhere (Barry & Friedman, 1998; Bechtoldt, De Dreu, Nijstad, & Zapf, 2010; Lyubomirsky & Nolen-Hoeksema, 1995). Teams whose members score lower on perspective taking, in contrast, may perceive negative affect, but devote less energy to understanding it, thereby allowing them to focus on their task. We therefore propose the following hypothesis:
We do not expect perspective taking to moderate the relationship between positive state affect and team performance. Positive state affect tends to be positively related to team performance (Knight & Eisenkraft, 2015). As positive affect should not consume fellow team members’ cognitive resources needed for the task at hand, there are no theoretical reasons to predict that perspective taking should moderate this relationship. To provide a comprehensive picture, however, we also examine the relationship between team positive state affect and perspective taking on team performance.
Method
Participants and Procedure
Participants were 196 undergraduate students at a large university in the Netherlands divided into 49 four-person teams.
1
More than one third (38.5%) of the participants were male and ages ranged between 17 and 48 years (
The study took 3 hr to complete and was part of a larger study on team processes and performance. 2 At the start of the study, participants completed a set of questionnaires for 15 min including the dispositional perspective taking measure, followed by an intensive, collectively administered 60-min training during which all elements of the task were explained and practiced together with a research assistant. This was followed by a 5-min break and a discussion about the strategy for 10 min. Immediately after the strategy discussion, teams engaged in the focal task for 30 min. Finally, participants completed another set of questionnaires, in which the first set of questions asked about their state affect during the task. Team members were seated in the same room for the duration of the task and data collection. Direct communication between the team members was possible during the period for strategy discussion and during task completion.
Task
Participants engaged in the Michigan State University–Distributed Dynamic Decision Making Task (MSU-DDD), which has been previously used in studies about team processes and performance (Beersma et al., 2003; Hollenbeck et al., 2002). A brief description of the task is provided below (see Beersma et al., 2003, for an extensive overview).
In the MSU-DDD task, participants monitor a grid depicted on their computer screens that is divided into quadrants. Each player has unique responsibility for one quadrant in which his or her base is located. Each player also has four different vehicles at his or her disposal (i.e., jets, helicopters, tanks, and Airborne Warning and Control System [AWAS] radar planes), enabling each player to engage in the full range of behaviors needed to participate in the task. Around the base and vehicles are several rings that players can use to detect, identify, and destroy targets. The detection ring has a radius of roughly six grid units and is used to monitor the task grid. The identification ring is smaller, roughly four grid units, and enables a team member to determine whether a target is friendly or unfriendly. To identify any target outside of the ring around the base, the player can launch one of its vehicles or ask team members to share their information. The four vehicles differ on four dimensions: range of vision, speed of movement, duration of operability, and weapons capacity. The base as well as one of the vehicles is only able to detect and identify targets, but three of the four vehicles have an extra ring, located between the detection and identification ring, that allows them to attack targets as well. Targets can be friendly and unfriendly enemy targets; friendly targets should not be attacked, whereas unfriendly targets should be attacked. In addition, targets vary in speed of movement.
Apart from the different quadrants, vehicles, and targets in the task, the task grid itself is divided into three zones: a neutral zone, a restricted zone, and a highly restricted zone. The restricted zone is indicated by a green square and the smaller, highly restricted zone is indicated by a red square. Targets can be identified in the neutral zone, but all targets should be allowed a free passage in this zone. Teams that destroy friendly targets are committing errors of engagement. When enemy targets enter the restricted zone, one point is deducted from the team score for every second the target remains in this zone. However, if an enemy target enters into the highly restricted zone, 2 points are deducted from the team score for every second it is there. The objective of the team is to limit the time the enemy targets are in the restricted or highly restricted zones. Teams were able to observe their absolute score on the task, but they were unaware of how their team performance compared with other teams. As such, during the task, they could not infer whether their performance score was high or low.
Measures
We measured perspective taking with a nine-item scale (Davis, 1980, 1983) measuring dispositional perspective taking tendencies on a 7-point Likert-type scale (α = .75). A sample item is “I believe there are two sides to every question and I try to look at both of them.” A higher score on this scale reflects a greater habitual tendency to take perspective. We conceptualized perspective taking at the team level as a team composition variable and we conceptualized the team task as an additive task (Steiner, 1972). For additive tasks, the potential performance of the group increases when group members, on average, possess more of a relevant resource, or, in contrast, score lower on average on a variable that might stand in the way of group performance. Data for individual perspective taking tendencies were therefore averaged across team members to create the team-level construct (see for a similar approach Van Kleef et al., 2009).
Negative state affect was measured with a four-item scale that measured each team member’s anger, disappointment, fear, and sadness during the task (Barrett & Russell, 1998). All items were formulated according to the following logic: “During the task I felt (emotion).” The state affect was measured on a 7-point Likert-type scale where a higher score reflected more agreement with the item (α = .75).
Positive state affect was measured with two items that assessed each team member’s happiness and content (Barrett & Russell, 1998). The items were phrased similarly to those used for negative state affect and were measured on the same 7-point Likert-type scale (
Because team members interacted and experienced shared events, we expected affect to transfer between group members and hence conceptualized negative and positive state affect as direct consensus variables (see Barsade & Knight, 2015; Chan, 1998; Menges & Kilduff, 2015). Agreement statistics regarding affect within teams support this conceptualization (average
In the current study, we were interested in joint, interdependent team performance. In the MSU-DDD task, performance as an interdependent team is best reflected by the
Each team started the team simulation with a defensive score of 50,000 points. For each second an enemy target resides in the restricted zone, 1 point was lost, and for each second an enemy target resides in the highly restricted zone 2 points were lost. Team defensive scores ranged from 33,685 to 46,017 points.
Control variables
Team perspective taking is a team composition variable, or configural property of the team, and is defined as originating in, or emerging from “individual team members’ experiences, attitudes, perceptions, values, cognitions, or behaviors” (Klein & Kozlowski, 2000, p. 217). There are no theoretical or empirical reasons why individuals’ habitual perspective taking scores should converge in teams composed via random assignment. However, variability in perspective taking differs between teams in that teams with the same average perspective taking score might differ substantially as to the distribution of scores within the teams. Differences in dispersion of scores between teams might be related to team performance. Therefore, Klein and Kozlowski (2000) recommend controlling for dispersion within teams in such cases, a recommendation we follow in our analyses.
Furthermore, this study was part of a larger research project for which hypotheses and measures were developed separately (see Beersma, Homan, Van Kleef, & De Dreu, 2013). As part of this larger research project, the teams included in this study were randomly assigned to one of the cells in a 2 (regulatory focus: promotion vs. prevention) × 2 (reward structure: team vs. individual rewards) experimental design. The regulatory focus manipulation was administered during the training phase of the study, in which the training either emphasized safety and security needs in the prevention focus condition, or the training emphasized growth and advancement needs in the promotion focus condition. The reward structure condition followed Beersma et al. (2003) and Johnson and colleagues (2006). In the individual rewards condition, team members were told that they would win a cash prize of €10.00 (approximately US$12.50) if they were one of the top-12 performing participants. In the team rewards condition, teams were told that they would receive a cash prize of €40.00 (approximately US$50.00) if they were one of the top-three performing teams. In the section “Results,” we first present the tests of the hypothesis that was central to the current article, and then, as a robustness check, present the same analyses but controlling for the variables relevant to the larger research project: regulatory focus, reward structure, and their interaction.
Results
Table 1 shows the means, standard deviations, and correlations of the variables of interest in this study. We predicted an interaction between team members’ tendency to take others’ perspective and the prevalence of negative state affect within teams, such that in teams in which members on average score higher on perspective taking, we expected a stronger negative relationship between team negative state affect and team performance. Hierarchical regression analyses were used to test this hypothesis, using separate analyses to examine effects of negative and positive affect to avoid multicolinearity (positive and negative state affect were significantly negatively correlated,
Means, Standard Deviations, and Correlations of the Variables of Interest (
We report unstandardized values in this table to allow for interpretation of values in light of scale ranges.
Prevention coded as −1, promotion coded as 1.
Individual rewards coded as −1, team rewards coded as 1.
*
Perspective Taking as Moderator of the Relation Between Negative and Positive State Affect and Team Performance.
*
Robustness Analysis Controlling for Manipulated Conditions.
Model 1 in Table 2 shows that although there were no main effects for team negative affect or team perspective taking, the interaction between negative state affect and perspective taking predicting team performance was significant and in the predicted direction (
The moderating effect of perspective taking on the relationship between negative state affect and team performance (
With regard to positive state affect, the interaction between perspective taking and positive state affect was unrelated to team performance (
Models 1 and 2 in Table 3 depict the results for our analyses when including experimental conditions as control variables (i.e., regulatory focus and reward structure). Although we made no predictions as to how these variables would affect the relationships of interest here (between team perspective taking and team affect and performance), it is important to demonstrate the robustness of our findings by controlling for the effects of the experimental conditions teams were in. We did so by testing a hierarchical regression model in which the regulatory focus condition, reward condition, and their interaction were included as additional control variables (see Table 3, Models 1 and 2). The hypothesized interaction between negative state affect and perspective taking on team performance was still significant in a one-tailed test (
Discussion
Whereas the positive relationship between state positive affect in teams and team performance is well established, research shows that the relationship between state negative affect and performance is not very clear and subject to the influence of moderating factors (Barsade & Knight, 2015; Knight & Eisenkraft, 2015). We proposed and found that team perspective taking is such a moderating factor, and argued that high perspective taking in teams with high negative affect may hamper performance because it distracts from focus on the task.
In contrast to previous studies demonstrating the positive effects of perspective taking, the current study thus sheds light on the “dark side” (the negative consequences) of perspective taking. Whereas previous research showed that understanding others relates to smoother social interactions (Davis, 1983), less stereotyping of out-group members (Galinsky & Moskowitz, 2000), more win–win agreements in negotiations (Galinsky, Maddux, et al., 2008), and more creative solutions in teams (Hoever et al., 2012), our study showed that greater perspective taking tendencies at the team level can also distract from performing well. Specifically, negative state affect of team members was negatively related to team performance when perspective taking in the team was high. Whereas teams that scored low, on average, on perspective taking, hardly seemed to be affected by negative affect in their team, the performance of high perspective taking teams, in contrast, suffered when they experienced high negative affect. This might be due to the latter teams becoming distracted by making efforts to understand each other’s affect. These efforts may pay off when team members’ attentional focus should be internal, as in negotiations or with creativity assignments, where understanding others’ emotional states is crucial for task accomplishment (Galinsky, Maddux, et al., 2008; Hoever et al., 2012). However, our results show that these efforts may be a liability when the task at hand requires team members to expend their cognitive resources on their external environment. Theories on perspective taking should take this into account by acknowledging that team functioning can also be hurt by perspective, taking and that the affective states of team members are important boundary conditions of the impact of perspective taking.
Interestingly, the findings of our study resonate with earlier research demonstrating that teams whose members were highly capable of recognizing the emotional states of other team members were more likely to make dysfunctional relationship-oriented (rather than more functional task-related) attributions when experiencing conflicts within the team (Bechtoldt, Beersma, Rohrmann, & Sanchez-Burks, 2013). Emotion recognition has also been related to increases in stress reactivity in the form of cortisol secretion when facing others’ nonverbal negative signals (Bechtoldt & Schneider, 2016). Together, the findings of these studies show that recognizing emotions and putting oneself in the shoes of one’s team members is not always functional for interpersonal processes and team performance.
Additional analyses revealed that positive state affect was positively related to performance regardless of the level of perspective taking in the team. This finding is consistent with prior research on the effect of positive state affect on both individual and team performance (Kaplan, Bradley, Luchman, & Haynes, 2009). We found no moderating effect of perspective taking on the relationship between positive state affect and team performance. Therefore, not only did perspective taking exacerbate the negative consequences of negative state affect, it also failed to help teams benefit more from their positive state affect.
Limitations and Possible Directions for Future Research
There are certainly limits to the conclusions one can draw from a single study with a small sample size of 49 teams. One issue that is important to examine further is what the boundary conditions are for the findings of the current study. One such boundary condition might be the nature of the task a team has to perform. Task interdependence has been demonstrated to have a significant impact on team processes and outcomes (e.g., Gully, Devine, & Whitney, 1995; Gully, Incalcaterra, Joshi, & Beaubien, 2002; Wageman, 1995). In tasks with a low degree of interdependence, team members are able to perform task-relevant activities by themselves with little input from other team members (Wageman, 1995). The effects in our study pertained to team performance requiring team members’ coordination, and we found no interactive effects of team affect and perspective taking on a performance dimension for which coordination was less important (see Note 3). Future research could benefit from better understanding the effects of perspective taking and emotion recognition under different conditions of task interdependence.
A different issue related to task context is that performance is operationalized in the domain of losses (i.e., teams lose points when enemy targets reside in the restricted zones). As a loss framing may have increased the vigilance of team members in attending to their fellow team members affective states, this could have exacerbated the effects of negative affect and perspective taking. Whether this is the case is obviously an empirical question, and it would be interesting to examine whether our effects can be replicated in tasks situated in the domain of gains.
Furthermore, the level of negative affect experienced by teams could constitute a boundary condition for the effects of perspective taking. The average level of negative affect was relatively low, and arguably our study therefore offers a conservative test of the influence of perspective taking. It is important that future studies find out if this is the case.
Another issue that future studies could examine is specifying the exact mediating variables responsible for the relationship between group affect and performance. In their meta-analysis, Knight and Eisenkraft (2015) examine social integration as a mechanism that partially mediates this relationship. The authors define social integration as an umbrella construct that comprises aspects of how people are positively linked to one another and to a group, including cohesion, identification and interpersonal attraction (Dineen, Noe, Shaw, Duffy, & Wiethoff, 2007; Knight & Eisenkraft, 2015; Smith et al., 1994). However, other cognitions and motivations may partially mediate this relationship as well. As noted by Knight and Eisenkraft (2015), more precise knowledge on why group affect and performance are related can foster the development of interventions that help groups cope with the consequences of negative affect.
Besides the exact mediators responsible for the relationship between group affect and performance, the causal direction of this relationship is also an issue that has been debated in the literature on group affect (Knight & Eisenkraft, 2015). A limitation of our study is that because we measured, rather than manipulated, state affect and perspective taking, it is not possible to specify the causal direction of the findings. Our prediction is based upon affect preceding performance, but it could be argued that state affect emerged as a function of how the group performed. To examine this possibility, we ran analyses testing for reverse causation. In these analyses, we found no evidence for an interaction between perspective taking mean and performance on negative affect. Therefore, the effects of perspective taking seem to be based on team members internalizing the team’s negative affect as if it was their own, rather than on taking the perspective of team members performing poorly.
Notwithstanding this finding, which rules out reverse causation in the current study, the issue of the causal relation between group affect and performance is of interest to the broader group affect literature. Knight and Eisenkraft (2015) argued that the relationship is likely bidirectional and reciprocal. Furthermore, they state that to address questions related to the direction of causality, studies that track the interplay of group affect, social integration and performance over time are needed, and we reiterate this recommendation. In addition, to strengthen confidence in the proposed causal relations specified by our theory, it would be interesting if follow-up studies would manipulate state affect and perspective taking (for examples of perspective taking manipulations used in a different setting see Galinsky, Maddux, et al., 2008; Hoever et al., 2012).
Finally, it is important to note here that we measured the habitual tendency to take perspective, but not the accuracy of perspective taking (i.e., whether high perspective takers correctly perceived team affective states). The pattern of results we found shows that the relationship between negative affect and performance is stronger for teams high in perspective taking, making it likely that high perspective takers accurately perceived their team members’ negative affect. If high perspective takers misperceived it for positive affect, we would not have observed this pattern of findings. However, it is also possible that high perspective takers inaccurately perceived negative affect as more negative than it actually was. It would be interesting to examine this in future studies.
Practical Implications
Organizations should take the nature of the task into account when composing teams. Composing a team of high perspective takers is not necessarily beneficial. For jobs in which team members need to focus on information in their external environment, such as military teams, air traffic controllers, or medical emergency teams, taking fellow team members’ perspective may backfire when they experience negative feelings. Although we measured perspective taking as a habitual tendency, other studies have found effects of perspective taking as a situational variable (Galinsky, Maddux, et al., 2008). This could mean that perspective taking has similarly negative implications when managers and team leaders encourage team members to understand each other. Thus, they should take care not to encourage team members to take perspective when tasks also require processing of information that is external to the team, and when the team experiences negative affective states.
Over time, a complete lack of perspective taking might interfere with team development. A possible solution for managers could be to build in phases of reflection in between task performance episodes, where there is room for emotional venting, taking team members’ perspective, and reflecting on team performance (Marks, Mathieu, & Zaccaro, 2001). If phases of intensive performance are intertwined with such reflection episodes, teams might reap the benefits of perspective taking without suffering from its distracting influence on performance (Schippers, Den Hartog, Koopman, & Wienk, 2003). In addition, our findings corroborate earlier research, which demonstrates that positive affect is positively associated with performance. Thus, leaders should not just take care to avoid negative affect in teams, but also focus on stimulating positive affect for optimal performance.
The current study demonstrated that perspective taking is not always an asset for teams; perspective taking negatively relates to performance in teams with negative state affect. Specifically, in tasks requiring constant processing of environmental information, team members’ mutual perspective taking results in worse performance when others feel negative. Teams low in perspective taking, whose members are less likely to pick up each other’s negative affective states, perform better. In other words, when considering the negative affect of one’s team members, ignorance may be bliss.