Synthesis Paper

Leader behaviors, conflict and member commitment to team-generated decisions

Igor Kotlyar a,1, Leonard Karakowsky b,⁎, Peggy Ng c,2

a Faculty of Business and Information Technology, University of Ontario Institute of Technology, Oshawa, Canada L1H 7K4 b School of Human Resource Management, York University, Toronto, Canada M3J 1P3 c School of Administrative Studies, York University, Toronto, Canada M3J 1P312

a r t i c l e i n f o a b s t r a c t

Available online 19 May 2011 Regardless of the quality of the decisions, teams that fail to effectively implement their decisions have accomplished nothing. Teams that derive decisions through a conflictual process may find that their members lack commitment to their decision due to the nature of the conflict the team experienced. This is where a team leader can play a critical role. The central aim of this laboratory study was to examine how leader behaviors can influence member commitment to team generated decisions via their impact on group conflict. Our study examined teams led by individuals who displayed charismatic-oriented behaviors, teams led by individuals who acted simply as pragmatic leaders, and teams whose members shared leadership. Consistent with our assertions, our findings indicate that pragmatic leaders were most effective at encouraging the highest levels of member commitment to team generated decisions through their restraint of dysfunctional forms of conflict during the decision-making process.

© 2011 Elsevier Inc. All rights reserved.

Keywords: Leadership Conflict Team Decision making

In participative decision-making groups, members have the freedom and the responsibility to contribute their own ideas and to challenge those of others during the decision-making process. The utility of team decision-making is not simply a function of the quality of the decision made. Effective team decision making also requires member commitment to the decision (Amason, 1996), where commitment reflects the degree to which members accept the decision reached and intend to cooperate in its implementation (Korsgaard, Schweiger, & Sapienza, 1995).

Decision making is successful only to the extent that the solution reached can be successfully put into practice. Group members who are not committed to their decision are unenthusiastic about seeing those decisions implemented. Lack of decision commitment can surface in the form of delays or even purposeful sabotage of implementation efforts (Guth & MacMillan, 1986). Although delays may seem to be a trivial concern, delays can constitute a potent tactic for killing an initiative (Pfeffer, 1992) and even slight delays can be detrimental in highly competitive and dynamic environments (Eisenhardt, Kahwajy, & Bourgeous, 1997). Consequently, team commitment to the implementation of the solution is of critical importance (Hitt & Tyler, 1991).

There is evidence that team conflict can play a powerful role in influencing commitment to team decisions. Specifically, relationship conflict (relationship conflict), which involves member disagreement over personal or emotional issues, generates member animosity and can detract from the quality of the decision process (Jehn, 1995; Mooney, Holahan, & Amason, 2007; Simons & Peterson, 2000). This, in turn, can reduce group members' commitment to group decisions (Janssen, Van de Vliert, & Veenstra, 1999). Regardless of the quality of the decision, members may simply not be committed to a decision that arose amidst a flurry of emotional conflict. Thus, the central question for group decision-making research is how group interaction should be

The Leadership Quarterly 22 (2011) 666–679

⁎ Corresponding author. Tel.: +1 416 736 2100. E-mail addresses: [email protected] (I. Kotlyar), [email protected] (L. Karakowsky), [email protected] (P. Ng).

1 Tel.: +1 905 721 8668. 2 Tel.: +1 416 736 2100.

1048-9843/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.leaqua.2011.05.007

Contents lists available at ScienceDirect

The Leadership Quarterly

journal homepage: www.elsevier.com/locate/leaqua

managed during the decision-making process so that it simultaneously produces decisions of high quality and high commitment (Amason & Sapienza, 1997; Dooley & Fryxell, 1999).

Group leaders are in the best position to manage team conflict and, therefore, to improve the effectiveness of decision-making groups (Amason, Thompson, Hochwarter, & Harrison, 1995), as well as to ensure member commitment to team decisions. Our study investigates how team leader behaviors can ultimately impact member commitment to team-generated decisions. We wanted to explore the impact of charismatic-related behaviors or behaviors that aroused emotion in the decision-making process. We also observed the impact of leader behaviors that focused solely on setting ground rules for conflict—what we refer to as pragmatic leader behavior (Mumford, 2006; Mumford, Antes, Caughron, & Freidrich, 2008). Finally we examined teams where no specific leader was designated to lead the team through the decision-making process. The aim was to assess the kinds of leader behaviors that functioned to engage teams in group decision making while constraining conflict that could harm member commitment to the decision that the team generated. Below we provide the theoretical background for our research, present our hypotheses and report the details of our study along with the results.

1. Theoretical background and framework

The proper use of the decision-making process requires dialectical interaction or conflict among the members of the group. The literature recognizes a conceptual distinction between two types of conflict. Task conflict arises through the discussion and debate of member preferences or opinions regarding the tasks at hand. If managed properly, this kind of conflict can be potentially productive in the sense that it can encourage exploration of many possible solutions before a final decision is reached (Amason, 1996). On the other hand, relationship conflict is most typically counterproductive given that its focus is on people rather than ideas (Jehn, 1995). That is, disagreements become personalized which damages trust and communication among team members (Janssen et al., 1999). The challenge then for decision-making teams is to manage productive levels of task conflict without triggering relationship conflict (Mooney et al., 2007).

It is important to note that while the researched has delineated these two types of conflict and have generally concurred that emotional conflict is negative, less agreement has been reached on the nature of task conflict. For example, in their meta-analysis, De Dreu and Weingart (2003) found that both task conflict and relationship conflict were negatively associated with team performance. This is in line with scholars who have suggested that the initial benefits of conflict quickly diminish as it intensifies and reduces group cohesion (cf., Tekleab, Quigley, and Tesluk (2009). On the other hand, a second view espouses the positive side of task conflict. According to this view, task conflict has a positive association with group outcomes like cohesion via the exercise of voice in team decision making (cf., Tekleab et al., 2009). However, it is acknowledged that this relationship exists only when emotional conflict is constrained and does not spill over from task conflict (Amason, 1996).

Several researchers have suggested that task conflict can trigger relationship conflict when task disagreements are perceived as personal criticism or disrespect (Amason, 1996; Medina, Munduate, Dorado, Martínez, & Guerra, 2005). Relationship conflict can also arise as a result of behavioral manifestations of negative emotions. In other words, this occurs when, in response to aroused anger, individual group members engage in aggressive behavior, ranging from slightly cynical comments to direct personal attacks (Wall & Callister, 1995). Of course relationship conflict can escalate when anger and aggressive behavior are reciprocated (Brett, Shapiro, & Lytle, 1998). The question of concern to this study is—how might different styles of leader behavior differ in their ability to manage conflict and ultimately maximize commitment to team generated decisions?

1.1. Leadership behaviors, conflict and decision commitment

Charismatic leadership theories (Bass, 1985) propose that a leader can achieve superior performance by modifying followers' needs and values. One particular element associated with charismatic leader behavior relates to the role of emotion. The role of emotion is connected to the ability of charisma to make esteem needs salient among the group. Under a charismatic leader, the followers' self- concepts are linked to the leader, the group's task and the group. This has the effect of harnessing the motivational force of self-esteem, self-worth and self-enhancement which make followers' efforts and goals more meaningful (Shamir, Zakay, Breinin, & Popper, 2000).

In addition to engaging followers' self-concepts (Shamir, House, & Arthur, 1993), charismatic leaders can also impact followers through the emotional attachment they build with followers (Gardner & Avolio, 1998). Consistent with Bass's (1985) observation, while pragmatic leaders focus on the rational or exchange basis of the leader–follower relationship, charismatic leaders emphasize the emotional basis of this relationship.

In line with that aspect, behaviors are aimed at increasing the intrinsic value of follower effort and goal accomplishment. Specific behaviors include expressing high expectations for followers to raise their self-esteem as well as expressing positive emotions to motivate followers (cf., Johnson, 2008; Connelly, Gaddis, & Helton-Fauth, 2002). Such positive affect can be communicated through the content of the leader's speech as well as through nonverbal leader behavior (Ashkanasy & Tse, 2000) and can facilitate the spread of affect to followers through emotional contagion (Johnson, 2008).

Emotional contagion has been investigated as one method by which leader affect can impact follower affect (Johnson, 2008; Sy, Cote, & Saavedra, 2005). Emotional contagion has been defined as the automatic and unconscious transfer of emotion between individuals (Hatfield, Cacioppo, & Rapson, 1992) that can occur as a result of the target individual mimicking the expressions or behaviors of others and subsequently experiencing the emotion that is being mimicked (Chartrant & Bargh, 1999). A number of recent research studies have found evidence that emotional contagion is the conduit through which leader affect can impact follower affect. For example, in her recent field study, Johnson (2008) found that leader affect at work was associated with follower

667I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

affect via the mechanism of emotional contagion. Similarly, Sy et al. (2005) found that team leaders' moods influenced team members' moods and impacted the affective tone of the group.

The salience of the self-concept suggests that group members will not only fight harder for their positions, they are also more likely to interpret criticism of their positions as a form of a personal attack. Making the need for esteem salient will encourage participants to become more personally sensitive to criticism. A perceived threat to one's self-image is likely to trigger relationship conflict as the individual attempts to repair a wounded self-image (Lazarus, 1993). This is much less likely to happen under where leaders display more emotion-neutral behaviors. In a work team context, we suggest that a leader who largely acts as a pragmatic leader and does not engage esteem needs of his followers will consequently bring less risk of triggering relationship conflict (Kotlyar & Karakowsky, 2006).

Pragmatic leader behaviors, as opposed to the charismatic-related behavior mentioned above, avoid making member ego or self-image salient. Such behaviors are simply focused on assisting the team through the decision-making process though focusing on adherence to the rules of the game rather than on stimulating ego or emotions. This can reduce instances of aggression and hostility by discouraging such behaviors, reprimanding inappropriate, emotionally-driven outbreaks and by clearly specifying the rules of conduct (Brett et al., 1998). In addition, this reduces reciprocation of aggression. For example, the pragmatic leader focuses on reminding members of their superordinate goal(s) (Eisenhardt, 1999), their common group membership, and their mutual dependence on each other (Pruitt & Rubin, 1986). Given the nature and role of the pragmatic leader behaviors, group members are likely to perceive little need to escalate conflict since the leader's explicit job is to enforce fair rules of conduct.

Pragmatic leader behavior focuses on facilitating team discussions via the implementation of rules and guidance. Actions of a group leader that promote certain rules of conduct and manage expectations of group members regarding the process can serve to reduce the level of frustration in group interactions. For example, rules of conduct may include the following: listen to others' ideas respectfully, without interrupting or making cynical comments; if failed to see the essence of others' arguments, seek clarification but do not use judgmental or “set-up” questions; critique ideas, but do not criticize individuals; challenge others' assumptions, but do not challenge their integrity, intelligence, or motives (Tjosvold, 1993). In this situation, making group members aware of the source of their arousal (i.e., reducing the likelihood of misattribution) can reduce the intensity of the emotion they experience (Schwarz & Clore, 1983). A leader can do this by explaining to the group members that it is normal to experience a sense of frustration and physical arousal when engaged in a dialectical interaction.

Our study also considers shared leadership situations—i.e., decision-making teams who operate without a leader and examine the impact of this on conflict and commitment in those teams. These are teams that simply report to an external facilitator or a consultant/advisor who is largely limited to inviting group members to take responsibility for their actions (Wageman, 1997) and encouraging them to lead themselves (Manz & Sims, 1987). Those facilitators often act as a non-invasive resource for the team instead of becoming directly involved in the team's task (Cohen, Ledford, & Spreitzer, 1996; Morgeson, 2005). In our view, one consequence of this approach is that the shared leadership teams do not afford the same level of defense from relationship conflict as do pragmatic leader behaviors. Rather the team is guided by shared leadership and this kind of leadership may lead to difference consequences with regard to team conflict generated.

We predict that relationship conflict will likely rise to disproportionately high levels because of the lack of a formal leader's willingness to impose concrete procedures on the group. The lack of involvement on the part of an external facilitator can create uncertainty over procedural direction and, thus, can contribute an additional source of conflict over the appropriate procedure to be used (Jehn, Chadwick, & Thatscher, 1997). Conflict over procedures tends to be dysfunctional (Jehn et al., 1997), and it can be particularly unproductive when perceived uncertainty exists over procedures for using structured conflict techniques. Moreover, once relationship conflict is sparked, a self-managed group may not be well equipped for containing conflict (Langfred, 2007). There is no formally assigned leader to interfere in a conflict among group members, even in those situations where task conflict appears to degenerate into relationship conflict.

Consistent with the arguments above, we assert that pragmatic leader behaviors will tend to be more effective at restraining relationship conflict and its damaging effects on commitment to the team decisions compared to teams who have shared leadership. Similarly, we assert that pragmatic leader behaviors are also more capable of restraining relationship conflict compared to charismatic leader behaviors—but for different reasons.

One can argue that the central features of pragmatic leader behaviors, including the establishment and reinforcement of explicit conflict generation goals, strengthen its management of relationship conflict in the team. Pragmatic leader behaviors, compared to charismatic-oriented and shared-leader behaviors, will likely be more effective at disconnecting the junctures at which task conflict can be transmitted into relationship conflict. The ego-engaging features of charismatic leader behaviors, and the “hands-off” aspect of shared leader behaviors, suggest that both these styles are likely to achieve less success in avoiding the triggers for relationship conflict. This in turn suggests that behaviors reflective of the pragmatic leader style will be more successful at attaining higher levels of member commitment to team generated decisions. Since these teams have not suffered from the dysfunctional effects of relationship conflict, they are much more likely to support the decision that has been reached by group efforts. The above arguments can be summarized in the following two hypotheses:

Hypothesis 1. Pragmatic leader behavior will generate lower levels of relationship conflict compared to behaviors reflective of charismatic-oriented and shared leadership.

Hypothesis 2. Pragmatic leader behavior will generate higher levels of member commitment to a team generated decision compared to behaviors reflective of charismatic-oriented and shared leadership.

668 I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

2. Method

2.1. Sample and design

This study used 208 university students (65% of the subjects were women, 35% were men) from undergraduate business programs in two large North American universities. Subjects were, on average, 21.4 years of age (st. dev.=2.7) and had 1.5 years of work experience (st. dev.=2.2). To achieve a reasonable level of statistical power (approximately 0.80) (Borenstein & Cohen, 1988), 69 groups were used in this study (approximately 23 groups per leadership condition).

Students were randomly assigned to groups and each group was assigned a leader (a confederate, blind to the hypotheses). No students had previously worked together. All leaders were males to avoid potential confounds based on gender, given the findings which have indicated that women behave differently than men in positions of leadership (Eagly & Johnson, 1990) and are perceived differently by their followers (Javidan, Bemmels, Devine, & Dastmalchian, 1995). Most groups were comprised of 3 members, excluding the leader (the exceptions were one group had 2 members and two groups had 4 members). Actors playing leaders were trained to behave in accordance with the three leadership styles (i.e., pragmatic leader, charismatic-oriented, shared). This approach is similar to other experiments, including Howell and Frost (1989), and Kirkpatrick and Locke (1996).

2.2. Procedure

Each group was assigned a leader (an actor) and a discussion space. Actors were introduced to the group members as “group leaders” (in the case of shared leaders, actors were introduced as “facilitators”), who were there to assist the groups in arriving at their decisions. Following the introduction, all groups were guided through a quick exercise, designed to familiarize group members with each other. The purpose of this part of the exercise was to better acquaint members with each other by engaging them in a discussion and also to introduce them to the dialectical approach to decision making.

2.2.1. Introduction of structured-conflict methodology Upon meeting his group, each leader introduced group members to the structured-conflict methodology, adapted from

Schweiger, Sandberg, and Ragan (1986). The leader verbally explained the structured conflict approach and the tactics for generating task conflict and minimizing relationship conflict to group members and clarified all points of uncertainty or misunderstanding (similar to Amason et al., 1995; Eisenhardt et al., 1997). He also distributed handouts outlining the tactics for generating task conflict and minimizing relationship conflict.

2.3. Task

Following the introduction, each group leader distributed a short case (approximately 2 pages in length) describing a real-life challenge faced by a company. One of two cases was used in each group. One case described a strategic issue that involved a competitive threat in the form of technological advancements. The other case required subjects to decide how to deal with complaints from customers regarding two employees and which involved personal value judgments and had ethical implications. No significant differences were found between these two cases with respect to task conflict (F2,63=0.48; n.s.), relationship conflict (F2,63=0.74; n.s.), or decision quality (F2,63=0.59; n.s.).

All participants received the same instructions: (a) identify the main problem(s) in the case, (b) recommend a solution(s) (which should be stated in terms of actions to be taken (based on Maier, 1963) and (c) logically present the supporting arguments and explicitly state all assumptions (based on Thomas, 1999). Each group was allotted 1 hour and 20 minutes to analyze the case in their groups and to prepare written recommendations.

In the pragmatic leader and the charismatic conditions, leaders led their groups through the steps outlined in the handout and then allowed their groups to work independently in order to arrive at consensus for their recommendations. In the shared leader condition, an external advisor or facilitator simply informed the group that there were to decide on the extent to which members wanted to follow the recommended steps. In all conditions, group members understood that any leader's job was to assist the groups in arriving at a high quality decision by guiding them through the process and not to provide advice on the substance of the case. Following the study, subjects completed a questionnaire which assessed their experiences in the group discussion.

2.4. Leader behaviors

Our aim was to examine whether certain types of behaviors commonly associated with the differing leadership behaviors could differentially impact task conflict and relationship conflict in the group. As our independent variable, we attempted to model behaviors central to and reflective of pragmatic leader behavior, charismatic-oriented leader behavior and shared leader behavior. There is evidence that leadership behaviors can be trained, including the behaviors suggestive of a charismatic style (Barling, Weber, & Kelloway, 1996; Kirkpatrick & Locke, 1996). Six male actors were hired and trained to play the role of group leader (similar to Howell & Frost, 1989; Kirkpatrick & Locke, 1996)—two actors for each behavioral leadership style. All leaders were blind to the hypotheses of the experiment. We trained the actors in the appropriate behaviors and practiced with them applications of these behaviors. The actors were required to learn a script for each leadership role and we assisted the actors in rehearsing the delivery of their lines. Together with the actors, we modified the script to ensure that the actors were comfortable with their lines

669I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

and that their delivery projected the intended leadership style. As discussed below (manipulation check), we attempted to ensure that the actors were playing their roles consistently and accurately for this study.

Prior to the experiment, a pilot study was completed in order to test and fine-tune operationalizations of pragmatic leader and charismatic-oriented behaviors (similar to Kirkpatrick & Locke, 1996). Leaders were told that their job was to help each of their groups do its best by stimulating critical analysis through debate and maintaining healthy inter-personal relations (i.e., minimizing relationship conflict). During training, three areas of the leader's job were emphasized: motivating group members to win the case competition, guiding group members through the structured-conflict process and encouraging individual group members to engage in the debate of ideas.

2.4.1. Pragmatic leader condition Each actor trained in the pragmatic leader leadership style was taught to display certain behaviors. To begin, he clarified the

group's goal (e.g., “You want to generate recommendations of the highest quality possible and win this competition”). He also encouraged each group to engage in task conflict by familiarizing group members with the structured-conflict process, assigning roles to individual group members, and leading the group through the process (i.e., keeping track of time, preventing any one group member from monopolizing the discussion, etc.). The leader repeatedly emphasized the value of the monetary reward for the best case solution and used praise and negative feedback to steer the group toward debate and away from relationship conflict. He also maintained positive interaction among group members by specifying acceptable and unacceptable rules of conduct, managing expectations, etc. Finally, the leader communicated to group members high performance expectations and exhibited confidence in their ability to meet these expectations (similar to Kirkpatrick & Locke, 1996). No significant difference was found between the two actors trained for this role with respect to the amount of task conflict (t=0.55; n.s.), relationship conflict (t=0.29; n.s.) and perceived leadership (t=1.28; n.s.).

2.4.2. Charismatic leader condition Actors trained to display charismatic qualities were taught to exhibit the same behaviors as those trained in the pragmatic

leader leadership style, with two exceptions. The first exception was that instead of repeatedly emphasizing the value of the monetary prize, these leaders appealed to the followers' higher order needs (i.e., achievement and esteem needs; e.g., “I want you to win this case competition so you can brag about it to your friends!”) and attempted to shift the perspective away from the concern with monetary reward to a grander perspective involving implications for self-identity and consideration of one's contribution to the group (e.g., “It's not about money; it's about pride.”). Based on Howell and Frost's (1989) methodology, actors playing the role of a charismatic leader introduced the case to subjects by emphasizing the value of the individual group members in achieving success. The second exception was that these actors were taught to demonstrate their enthusiasm and energy through their dynamic presence (House, 1998). They achieved this through voice intonation and body movements (Howell & Frost, 1989). The charismatic-oriented leader's behavior was operationalized by having the actors alternate between pacing and sitting on the edge of the desk, leaning toward subjects, maintaining direct eye contact when speaking, varying the rate of speech and loudness of voice, using hand gestures for emphasis when speaking, and having a relaxed posture and displaying animated facial expressions (Howell & Frost, 1989; Kirkpatrick & Locke, 1996). No significant difference was found between the two actors trained in charismatic-oriented behaviors with respect to the amount of task conflict (t=0.88; n.s.), relationship conflict (t=1.16; n.s.) and perceived leadership (t=0.17; n.s.).

2.4.3. Shared leadership condition External facilitators served primarily an advisory role. Each facilitator communicated to his group (a) that the group had the

power to decide on the approach to working as a team and to make its own recommendations and (b) that the role of the facilitator was only that of an advisor. The facilitators recommended the structured-conflict approach, but also advised groups that they had the authority to decide on any process that they felt would be appropriate. Facilitators carefully explained the structured conflict process and provided groups with a list of steps. They also warned group members about the dangers of relationship conflict, told them how to prevent relationship conflict, and provided them with a list of actions they could take to minimize relationship conflict. Facilitators also encouraged group members to praise each other for good performance, to discourage poor performance, and to set high expectations for themselves (as per Manz & Sims, 1987). When groups began reading the case, the facilitator typically left the room and let groups work on their own. No significant difference was found between the two individuals playing the external facilitator role with respect to task conflict (t=1.00; n.s.), relationship conflict (t=0.55; n.s.) and perceived leadership (t=0.08; n.s.).

2.5. Measures

2.5.1. Task conflict This variable was measured with a 4-item scale, with the end points anchored by “none” (1) and “a great deal” (7). The

structure of the questions was similar to those used by other researchers (Amason, 1996; Amason & Sapienza, 1997; Jehn, 1994), including “In analyzing this case, how many differences of opinion over content did your group have to work through?; How much disagreement in terms of ideas/opinions was there in your group at the beginning of your group discussion?; How much disagreement in terms of ideas/opinions was there in your group at the middle of your group discussion?; In arriving at the final

670 I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

recommendations, how much disagreement over different ideas/opinions was there?”. The 4-item measure of task conflict had a coefficient alpha of 0.85.

2.5.2. Relationship conflict This variable was measured by using a three-item scale (Amason, 1996; Jehn, 1995): “How much anger was there among the

group members during this decision?” “How much personal tension was there in the group during this decision?” “How evident were personality clashes between group members during this decision?” The coefficient alpha of this scale was 0.90.

2.5.3. Team decision commitment Commitment to an identified strategic decision was measured by a modified 6-item scale, using a 7-point Likert-type scale,

based on Dooley and Fryxell (1999) and an additional item from Schweiger et al. (1986; Priem, Harrison, & Muir, 1995). The items measured the extent to which team members were willing to stick with the decision through implementation and exert effort to see it implemented. We used the following items: “How risky is it for the company depicted in the case to follow your group's recommendations?” (reverse-coded), “How willing would you be to defend your group's recommendations in front of the executives of the real-life company depicted in the case?” “Is it likely that obtaining more detailed information about the situation depicted in the case may cause you to reduce your support for your group's recommendations?” (reverse-coded), “How much do you care about seeing your group's recommendations successfully implemented?” “How proud are you to tell others that you were involved in making these recommendations?” “How committed are you to your group's recommendations?” and “If your schedule permits, how willing would you be to participate in the implementation of your group's recommendations?” Coefficient alpha for this 7-item scale was 0.70.

2.5.4. Perceived quality of decision outcome (team measure) Perceived quality of outcome was measured with 6 items, using a 7-point Likert-type scale based on Amason (1996). Subjects

were asked the following questions: “Did your group generate recommendations of high quality?” “How well did your group diagnose the problem in the case?” “How likely are your group's recommendations to win this competition with other groups?” “Did your group conduct a thorough analysis of the case?” “How well did your group's recommendations address the problem presented in the case?” “If your recommendations were to be implemented by the organization, how likely are they to be successful in overcoming the problem presented in the case?” This 6-item scale had a coefficient alpha of 0.82.

2.5.5. Judged quality of decision outcome (independent rater measure) Group performance was also be assessed independently by two raters, who evaluated the quality of recommendations on a

scale ranging from 1 (low quality) to 7 (high quality) according to how well the group identified the problem(s) in the case and on the quality and logical support for the proposed solution (Priem et al., 1995; Schweiger et al., 1986; Schwenk & Cosier, 1993; Thomas, 1999). Two independent raters were used—a PhD graduate and a PhD student, both experienced in evaluating case studies, to assign scores. The raters were blinded to the experimental condition of each group. Judges were instructed to follow a process similar to that used by Priem et al. (1995). First, judges were required to separate papers into three piles of roughly equal size as they read them: poor, average, and good/excellent. Second, they were to re-read the papers and order them in terms of quality within each pile. Third, judges were to review the papers and continue re-ordering them until they felt confident in their evaluation. Lastly, they were to assign ratings. The average of the two judges' scores was used for analysis. The inter-rater correlation was 0.66. Other researchers who have used this approach reported inter-rater correlations of between 0.64 (Schweiger et al., 1989) and 0.95 (Priem et al., 1995).

2.6. Control variables

Before subjects engaged in case analysis as a group, subjects' expectations of conflict were measured. It was anticipated that expectations of conflict could contribute to cognitive conflict and relationship conflict. For example, high expectations that affective conflict would flame up could have suppressed group members' willingness to assert their opinions, particularly if they felt highly uncomfortable experiencing interpersonal conflict. High expectations of conflict could have also materialized into higher levels of relationship conflict, because group members who held such expectations would have been more likely to (mis) interpret communication from other group members as being hostile toward them and may have reacted accordingly. The following items were used to measure expectations of conflict: “How well do you think you'll be able to work with the other individuals in your group?” “How much interpersonal friction do you expect among the group members during the decision- making process?” “How many personality clashes do you expect to see in your group as it works toward its recommendations?” The coefficient alpha for this scale was 0.65.

Upon completion of the case study, subjects were asked to indicate their age and sex. It is feasible that differences among group members in terms of these elements could have impacted cognitive conflict (e.g., younger group members may not have felt that it was their place to assert their views in the presence of older group members). Such differences can also constitute an additional source of interpersonal friction and, thus, could have potentially contributed to relationship conflict. For the purpose of analysis, individual measures of age and sex were converted into measures of dissimilarity for each group by using the formula for Euclidian distance.

671I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

Subjects were also asked to report their first language (i.e., English, Chinese, etc.). Number of different first languages in each group was used as a proxy of ethnical/cultural diversity. It is quite feasible that cultural ethnical/cultural diversity within a group could have resulted in higher cognitive conflict (say, due to more varied opinions and different cultural perspectives on any one topic) and higher relationship conflict (say, due to higher likelihood of miscommunication and misunderstanding of intent).

Subjects were also requested to complete a shortened version of a social skills measure (similar to Riggio, 1986; Sessa, 1996). The scale produced an extremely low coefficient alpha and, consequently, it was not used in the analysis.

3. Results and discussion

3.1. Aggregation

We collected the data at the individual-level and aggregated it to the group-level by taking an average of individual scores. To justify aggregation of the studied variables, we computed the eta-squared statistic for the variables used in the analysis. Eta- squared indicates whether people within the group are more similar than people in different groups. Eta-squared values for perceived quality, task conflict, relationship conflict are 0.36, 0.54, 0.50, respectively. These values exceed the minimum recommended criterion of 0.20 (Georgopolous, 1986) and are consistent with those used in other studies (e.g., Simons & Peterson, 2000). Consequently, aggregation to the group level was justifiable.

3.2. Group adoption of structured conflict process

We compared the three leadership conditions in terms of the extent to which groups followed the steps outlined in the written instructions for generating task conflict that were provided to them by their group leader. Analysis of this one-item, 7-point Likert scale, self-report measure (“In arriving at your recommendations, did your group follow the steps outlined in the "Instructions for generating productive conflict?”), suggests that there is no significant difference between pragmatic leader and charismatic conditions (t=0.71; n.s.). However, groups working in the shared leadership condition (with external facilitators) indicated that they followed the instructions to a lesser degree (M=5.00, std=1.00) than those working with leaders who exhibited pragmatic leader leadership behaviors (M=5.63, std=0.63) (t=2.79; pb0.05) or charismatic-oriented leadership behaviors (M=5.80, std=0.66) (t=3.45; pb0.01).

3.3. Covariates

Before proceeding with the analysis, we considered whether the control variables should have been used as covariates in the analysis. We compared each of these variables across the experimental conditions and found no significant difference in terms of expectations of conflict (F5,63=1.41; n.s.), age dissimilarity (F5,63=1.57; n.s.), sex dissimilarity (F5,63=1.85; n.s.), and number of languages (F5,63=0.22; n.s.). Consequently, we did not include any covariates in subsequent analysis.

3.4. Manipulation check

As described above, in this study, we manipulated three styles of leadership behavior. To help assess face validity of our manipulations, we conducted a pilot study to determine the actors' ability to stay “in character”. From behind a two-way mirror, we observed how the actors led groups of subjects through the process from start to finish. We were satisfied that the actors adhered to the script and were able to properly exhibit the assigned leadership behaviors. To check whether the actors continued to exhibit the appropriate behaviors during the experiment, we conducted random spot checks. We periodically observed and listened to groups' deliberations, while trying to make our presence as unobtrusive as possible, so as not to affect the group dynamics. These random checks indicated that the actors' behaviors were being portrayed in a manner consistent with their assigned roles.

In addition to our periodic spot checks, we attempted to measure subjects' perceptions of the leaders (similar to Kirkpatrick & Locke, 1996). Following the exercise, subjects answered the following three questions on the Likert-type 7-point scale: “Did the individual guiding your group heighten your personal motivation to do well in this case competition?”, “Did the individual guiding your group motivate you personally to do better than you originally expected you would do?”, and “Did the individual guiding your group inspire you to assert your personal ideas and opinions?” The alpha coefficient for this “perceived leadership” scale was 0.82.

We conducted an analysis of variance to examine the differences among the three behavioral styles with respect to “perceived leadership.” Although the pattern of means was in the expected direction, there was no significant difference among the three leadership conditions (F2,66=2.36; n.s.). However, the absence of a significant difference should not be interpreted to mean that the leadership behavior manipulation was not successful. First, observations during the pilot study and the occasional spot checks during the experiment indicate that the actors adhered to the script and projected behaviors consistent with their assigned roles. Second, there are potential explanations available for the lack of significant differences, which suggest that the problem is likely due to the poor procedure for a manipulation check, rather than the absence of a manipulation.

To begin with, the questions that we used for the manipulation check were aimed to gauge the change in the subjects' motivation, however, the subjects themselves would not necessarily notice the change because leaders may “arouse follower nonconscious motives” (House & Shamir, 1993: 91). In discussing the impact of a transformational leader, House and Shamir

672 I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

(1993: 92) argue that “individuals may or may not even be aware of the specific stimuli that aroused their motives or the specific behaviors that are mobilized by such stimuli.” The problem may have been made worse by the way in which we administered these questions. Subjects in each leadership condition were not exposed to the other two styles of leadership and, consequently, they could not compare “leadership” of their group leader with the other conditions. Instead, subjects might have quickly categorized their group leader and compared him against that mental stereotype (cf. Phillips & Lord, 1982). The absence of a contrast could have suppressed the difference among the groups and produced non-significant results.

Clearly this is a deficiency that reduces the strength of our study. However, it is instructive to note that other researchers, who have manipulated leadership behavior using a similar approach, have also reported that subordinates did not perceive the leadership manipulations as different, but their manipulations caused differences in subordinates' performance (e.g., Peterson, 1997). This was also the case for our study—i.e., below we do report different impacts of these leader behaviors on a number of variables of central interest in this study. In the future, a different form of manipulation check should be used, one that allows for a more accurate comparison among the leadership conditions. For example, it might be wise to video tape groups' deliberations and have external observers, blind to the conditions, independently view the tapes and rate all leaders in terms of their leadership style. Such comparison is likely to yield more accurate results.

3.4.1. Decision quality We did not predict any outcomes regarding decision quality. In fact, our view is that commitment to a decision can occur quite

apart from any perceptions of the quality of the team's decision. Nonetheless we wanted to test for any possible differences in judged or perceive quality of the decision outcome. We tested for differences in team member perceptions of decision quality based on their leader's style. We used ANOVA, treating perceived quality (team measure) as the dependent variable. Table 1 presents the means of the three types of leadership behavior. There are no significant differences among the three leadership conditions. The variance of perceived quality is 0.41.

With regard to our rater measure of decision quality (judged quality), we also used Levene's test to compare the variances across the three conditions of leadership behavior (Kirk, 1995). This indicated that there are no significant differences among leadership conditions with respect to judged quality (F2,66=0.23; n.s.) (see Table 2).

While there was no significant relationship found between leadership and decision quality, we can also consider the extent to which conflict may mediate the relationship between leadership style and performance. Based on previous research, to illustrate that conflict functions as a mediator between leadership style and performance, three conditions must be met (Baron & Kenny, 1986). First, leadership must significantly account for variation in CC and AC. Second, conflict must significantly account for variation in performance. Third, when conflict is introduced into the equation, the relationship between leadership style and performance must lose its significance.

We first examined the issue of mediation with respect to judged quality. The presence of the first condition has already been established: leadership style has been shown to have a direct effect both on CC and AC, as indicated above. To explore the second condition, we regressed judged quality on CC and AC. The results reveal significant coefficients for both CC (t=3.75; pb0.01) and AC (t=−3.56; pb0.01). Thus, the second condition is established. To explore the third condition, we employed hierarchical regression analysis. Transformational and facilitative behaviors were inputted into the model; transactional leadership behavior was treated as the reference category.

Results of the hierarchical regression model are displayed in Table 3. When conflict variables are introduced into the equation in Model 2, coefficients for leadership behavior remain significant (βtsfrm, t=−3.20; pb0.01, and βfacil, t=−2.16; pb0.01). This analysis indicates that conflict is a partial mediator of the relationship between leadership behavior and judged quality (independent rater measure). The leadership behaviors add explanatory power over and above that of conflict.

We also used a regression model to examine the same question regarding perceived quality (team measure). The results of the regression are presented in Table 4. Although perceptions of quality are not affected by leadership behavior, they are affected by conflict. Coefficients for both CC (t=2.02; pb0.05) and AC (t=−2.97; pb0.01) are significant. This suggests that subjects are interpreting the quality, in part, from the amount of issue-focused debate and, in part, from the amount of animosity that exists among the group members.

In sum then, conflict appears to function as a partial mediator between leadership behavior and performance. With respect to judged quality and commitment, leadership behaviors provide some explanatory power over and above that provided by conflict.

Leadership behaviors do not have an effect on perceived decision quality and subjects' perceptions appear to be formed on the basis of perceived conflict.

Table 1 Means and standard deviations of perceived decision quality by leader behavior.

Mean St. dev. N

t

1 2

1. Pragmatic leader 5.84 0.41 24 2. Charismatic 5.64 0.41 23 1.68 3. Shared 5.66 0.41 22 1.50 0.17

Note. Bonferroni adjustment for multiple comparisons. *pb0.05.

673I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

One observation that emerges from the analysis is that AC plays an important role in performance of decision-making groups. Previous research has focused primarily on the effects of cognitive conflict (e.g., Schweiger et al., 1986; Schwenk & Cosier, 1993). These results indicate that affective conflict maybe just as important for understanding performance of decision-making groups.

3.4.2. Leader behaviors and task conflict We did not propose any hypotheses regarding task conflict. However, it was necessary to determine the level of task conflict

each leader generated given its potential influence on relationship conflict. Data from the experiment found that charismatic- oriented behavior generated the highest levels of task conflict (F=7.00; pb0.01). The amount of task conflict generated by this type of leadership style is significantly higher than that generated by pragmatic leader behaviors (t=3.03; pb0.05) and facilitative behaviors (t=3.41; pb0.01). However, there is no statistical evidence of a difference between those of facilitator and pragmatic leader behaviors (t=0.46; n.s.) (see Table 5).

There may be a number of reasons why charismatic-oriented leader behaviors tended to generate more task conflict than the other two conditions. Given the requirement of introducing structured conflict methodology into the group process, it could be that this type of leader behavior was more successful in gaining the commitment of members to engage in the kind of brainstorming or constructive conflict necessary to generate potential decisions. Groups in conditions under shared leadership and pragmatic leaders may have been less capable of stimulating members to “dive in” in terms of tackling the problem at hand and engaging in actively thrashing out solutions. Clearly, more research is needed to delineate possible reasons for this difference.

3.4.3. Leader behaviors and relationship conflict Hypothesis 1 asserted that teams led by individuals who displayed pragmatic leader behaviors would generate lower levels of

relationship conflict compared to teams led by individuals who exhibited charismatic-oriented or who were in the shared leadership condition. To compare the effectiveness of leadership styles in their ability to minimize relationship conflict, analysis of variance technique was utilized. There is statistical evidence that pragmatic leader leadership styles generated significantly less relationship conflict than leaders who employ charismatic-oriented behaviors (F=13.54; pb0.01). Multiple comparisons on the means reveal that pragmatic leader behaviors are associated with lower levels of relationship conflict when compared to charismatic-oriented behaviors (t=5.14; pb0.01) (see Table 6). Given that levels of task conflict could influence the level of relationship conflict experienced, we conducted analysis of covariance to identify the type of leadership style that generates a less amount of relationship conflict conditional on any given level of task conflict. The model reveals that pragmatic leader behaviors led to a less amount of relationship conflict on the average (B=−0.43, pb0.05) than the charismatic-oriented behaviors. This supports the assertions of Hypothesis 1. However, it should be noted that while multiple comparisons indicate that pragmatic leader behaviors are associated with lower levels of relationship conflict when compared to teams with shared leadership, these differences are not statistically different (t=1.73; n.s.) (see Table 7). Consequently, in sum, Hypothesis 1 received partial support.

Table 2 Means and standard deviations of judged quality by leadership style.

Mean St. dev. N

t

1 2

1. Pragmatic 4.04 1.08 24 2. Charismatic 2.85 1.08 23 3.76 ⁎⁎

3. Shared 3.13 1.08 22 2.85 ⁎ 0.85

Note. Bonferroni adjustment for multiple comparisons. ⁎ pb0.05. ⁎⁎ pb0.01.

Table 3 Hierarchical regression analysis of judged quality.

Variable Model 1 Model 2

B T B B t B

Transformational −1.19 −3.76 ⁎⁎ −0.48 −1.08 −3.20 ⁎⁎ −0.43 Facilitator −0.91 −2.86 ⁎⁎ −0.36 −0.64 −2.16 ⁎ −0.25

CC 0.66 4.00 ⁎⁎ 0.48 AC −0.51 −2.91 ⁎⁎ −0.38 R2 0.19 0.37 ΔR2 – 0.18

F for ΔR2 (df) – 8.88 ⁎⁎

F for the model (df) 1 7.78 ⁎⁎ 9.26 ⁎⁎

⁎ pb0.05. ⁎⁎ pb0.01.

674 I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

3.4.4. Leader behavior and team decision commitment Hypothesis 2 asserted that teams led by individuals exhibiting pragmatic leader behaviors will show higher decision

commitment compared to teams led by charismatic-oriented style or in shared leadership teams. An omnibus F test shows that there are differences in decision commitment resulted among the leadership styles (F=3.69, pb0.05) which provides support for these assertions. We then considered the comparison between the individual leader styles. Based on the multiple comparison methods, statistical difference in decision commitment is significant between the pragmatic leadership condition and shared leadership condition (t=2.62, pb0.05). Once again, we controlled for levels of task conflict generated in the teams in order to assess whether leadership styles bring differences in team decision commitment after adjusting for the effect of task conflict. A linear regression model was fitted to assess this effect. The model showed a significant lesser amount of decision commitment generated under the shared leadership compared to groups under a pragmatic leadership style (B=−0.29, pb0.05). This is supportive of the assertions of Hypothesis 2 (see Table 8). However, the results indicated that there was no evidence of a significant difference in decision commitment in groups led by pragmatic leader leaders compared to teams led by charismatic- oriented leaders (though the means were in the predicted direction). Consequently, in sum, Hypothesis 2 received partial support.

The fact that charismatic leader behaviors were associated with higher levels of relationship conflict compared to pragmatic leaders did not seem to translate into significant differences in decision commitment between these two conditions. There may be a number of reasons why this difference was not significant. It could be that the nature of the impact of these charismatic leaders was such that group members' commitment while potentially “damaged” by relational conflict” may have been counterbalanced by the need to feel supportive of the charismatic leaders—that of commitment to the leader's goals. There may be a dialectic occurring between the detachment from a group fraught with conflict and the relationship with the leader. Consequently decision commitment may be a function of the relative power of these two forces. Clearly, future research needs to further assess the relative weight of these two forces.

Another possible reason for the lack of significant differences in decision commitment among charismatic and pragmatic leader conditions could stem from the protect ego and esteem needs. As discussed earlier, charismatic leaders typically make ego and identity needs salient. This suggests that among the cognitive consequences of such influence are attitudes towards the outcomes of team efforts. Perhaps through cognitive dissonance, team members believe that there are committed to a decision that they helped the team generate. This too will compete with the negative influence of relational conflict on their attitudes towards the results of their team's efforts.

4. Practical and theoretical implications

While numerous studies have considered the dynamics through which teams make decisions, few studies have considered what happens after the team arrives at their decision. This is certainly critical in determining the ultimate success or failure of the decisions. Decision process determines not only the quality of the decision but also the ultimate commitment exhibited toward the decision. The highest quality decisions can be considered failures if their implementation never occurs due to lack of team commitment to enforcing their decisions.

Table 4 Regression analysis for perceived quality.

Variable B t B

Transformational −0.06 −0.47 −0.07 Facilitator −0.09 −0.75 −0.10

CC 0.13 2.02 ⁎ 0.28 AC −0.21 −2.97 ⁎⁎ −0.44 R2 0.17

F for the model (df) 3.37 ⁎

⁎ pb0.05. ⁎⁎ pb0.01.

Table 5 Means and standard deviations of task conflict by leader behavior.

Mean St. dev. N

t

1 2

1. Pragmatic leader 3.26 0.80 24 2. Charismatic 3.97 0.80 23 3.03 ⁎

3. Shared 3.15 0.80 22 0.46 3.41 ⁎⁎

Note. Bonferroni adjustment for multiple comparisons. ⁎ pb0.05. ⁎⁎ pb0.01.

675I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

Our study attempted to examine the consequences on decision commitment within teams operating under charismatic, pragmatic or shared leadership. Our view is that the impact of leader behavior on commitment will come about largely through the influence on relationship conflict generated in the group. Our study offers support for our view that leader behaviors can contribute in important ways to enforcing or undermining team commitment to the decision—and this effect comes about largely through the leader's ability to manage conflict throughout the team decision process.

With regard to team conflict, the results indicated that behaviors reflective of pragmatic leaders tend to generate less relationship conflict compared to shared leadership and charismatic leadership. Lower levels of relationship conflict suggest that pragmatic leader behaviors are better at restraining such conflict to a greater extent than in the other conditions. This pattern of results is consistent with the view that self-managed teams can fail to respond adequately to conflict (Langfred, 2007). In addition, these findings also support the assertion that charismatic-oriented leaders may unintentionally generate higher levels of relationship conflict because of their potential for evoking self-esteem and thereby encouraging group members to become more emotionally involved in group conflict (Kotlyar & Karakowsky, 2006). In sum, groups led by pragmatic leaders may have a greater capacity for constraining some types of team conflict compared to groups with shared leadership or charismatic leaders.

We assert that the capacity to restrict conflict facilitates a higher level of decision commitment. Pragmatic leader leadership behaviors appeared to be associated with higher levels of decision commitment than the other leader behavior conditions. And this relationship appears to exist regardless of the actual quality of the team generated decision. That is, we found that, in all cases, there was no actual difference in perceived or judged quality of the teams' decision outcomes. This is interesting because it suggests that member commitment is more affected by team processes (i.e., the level of conflict present in the team dynamics) than the actual team outcome.

These findings have important practical implications. When using the “structured conflict” approach, leaders should be involved in the process by guiding group members through the decision-making process and should rely on extrinsic motivation to stimulate a debate of ideas. Leaders should be careful about challenging group members at a personal level or evoking self- concept. Triggering higher order needs in their followers can “over-stimulate” them and create relationship conflict. This advice is consistent with that offered by Janis (1982) that controversy should be introduced in a low-key style.

This study supports the assertion that pragmatic leader behaviors can minimize relationship conflict while permitting constructive task disagreement and discussion. However, there may be other costs associated with the kind of rules of conduct and expectation management that these leaders apply. It may be possible that some kinds of tasks require a level of frustration, perhaps even a level of emotional conflict that might be sacrificed when a pragmatic leader seeks to restrain conflict in this way. Future research needs to test these ideas under a diverse range of tasks.

In terms of theoretical implications, our study sheds additional light on the nature of the relationship between leadership behavior and group conflict and is intended to contribute to theory-building in this relatively uncharted area. However, it is important to note that this study did not include an actual measure of the specific processes through which leadership behavior results in relationship conflict or task conflict. For example, while we posited that encouragement of ego involvement may make the team more prone to relationship conflict, alternative explanations cannot be fully ruled out.

Table 6 Means and standard deviations of relationship conflict by leader behavior (task conflict is not included in the model).

Mean St. dev. N

t

1 2

1. Pragmatic leader 1.78 0.75 24 2. Charismatic 2.91 0.75 23 5.14 ⁎⁎

3. Shared 2.17 0.75 22 1.73 3.31 ⁎⁎

Note. Bonferroni adjustment for multiple comparisons. ⁎⁎ pb0.01.

Table 7 Means and standard deviations of relationship conflict by leader behavior (controlling for task conflict).

Mean St. dev. N

t

1 2

Pragmatic leader Charismatic

1. Pragmatic leader 1.86 0.69 24 2. Charismatic 2.71 0.73 23 3.95 ⁎⁎

3. Shared 2.29 0.70 22 2.12 1.86

Note. Bonferroni adjustment for multiple comparisons. ⁎⁎ pb0.01.

676 I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

5. Limitations and future research directions

It is important to point out several critical limitations of our study. First, as we acknowledged above, our findings are weak—and only partly supported. We endeavored to explore an area of leadership research that has largely been ignored. Our study serves largely as an initial attempt to discern possible differences in leadership behaviors on team conflict. Trying to simulate leadership behaviors, particularly those reflective of the charismatic style, is fraught with challenges.

Our study attempted to capture several leader behaviors that are fundamentally characteristic of three styles of leadership. However, our manipulation check failed to reveal significant differences among members in their reported perception of these behaviors. Consequently, our manipulation check was limited to a confirmation of the face validity of our manipulations. In the future, a different form of manipulation check should be used, and one that allows for an accurate comparison among the leadership conditions (e.g., video- tape groups' deliberations). In addition, even within the confines of our laboratory study it is difficult to isolate those behaviors that are solely the domain of each leadership style. For example, pragmatic leader leaders engage in forms of goal setting and the facilitation of self-efficacy, both of which are also reflected in the behaviors modeled by charismatic-oriented leaders. This is an area that future researchers must grapple with in efforts to move such research forward.

A number of other limitations of this study restrict the generalizability of our results. First, groups worked within a short-term, limited duration and were composed of undergraduate students as subjects who, compared to real-life decision makers, tend to have less experience in making complicated organizational decisions. However, it is difficult to predict how this limitation impacts the pattern of results. With respect to the role of conflict in decision making, students are commonly used for studying decision-making groups (Jehn & Mannix, 2001), and the results obtained in the laboratory tend to be similar to those obtained in the field (e.g., Amason, 1996). In addition to these groups being temporary in nature, the members also may have been influence by the fact that they likely would not work together in the future. Consequently, conflict could have also been mitigated by factors such as the expectation of working together in the future. It is also difficult to speculate as to how this may influence groups in our study. Knowing that they would not be working together in the future may have made individuals feel less inhibited from exhibiting aggressive behaviors. A second factor limiting the generalizability of our study is the use of only male leaders. Given the potentially important role of gender in group dynamics and leadership, it would be useful to replicate this research using female group leaders.

Our results indicated that not all groups equally adopted the structured conflict process. Specifically, as mentioned above, the teams operating under shared leadership apparently followed instructions to a lesser degree than the teams led by pragmatic leader and charismatic-oriented leaders. Future research needs to examine this discrepancy. While this raises some question regarding the interpretability of our results, it may be unavoidable. For example, the nature of the shared leadership groups could be such that these groups felt less guided compared to groups with a formally assigned leader (be it charismatic-oriented or pragmatic leader). The very nature of these roles could account for the differences in the degree of adoption of the conflict methodology. Consequently, some differences in conflict exhibited may be due to the failure to completely apply the conflict methodology, which is a secondary consequence of being in a group operating under shared leadership.

The central aim of the present study was to examine how leadership behaviors can influence the nature and level of conflict and commitment generated in the team context. Future studies should attempt to further discern the characteristics of the different leadership styles with regard to their differential impact on conflict. Future research is also needed to test the underlying mechanism or link between leader behavior and conflict. We did not directly test the potential drivers of this link—esteem needs, ego or identity. Consistent with Burke et al.'s (2006) observation, there is a need to develop a greater understanding of the link between leader behaviors and team performance outcomes, including behavioral or cognitive outcomes. It is our hope that this study will spark future research efforts aimed at more fully exploring these relationships.

References

Amason, A. C. (1996). Distinguishing the effects of functional and dysfunctional conflict on strategic decision making: Resolving a paradox for top management teams. Academy of Management Journal, 39(1), 123–148.

Amason, A. C., & Sapienza, H. J. (1997). The effects of top management team size and interaction norms on cognitive and affective conflict. Journal of Management, 23(4), 495–516.

Amason, A. C., Thompson, K. R., Hochwarter, W. A., & Harrison, A. W. (1995). Conflict: An important dimension in successful management teams. Organizational Dynamics, 23(2), 20.

Table 8 Means and standard deviations of decision commitment by leader behavior.

Mean St. dev. N

t

1 2

Pragmatic leader Charismatic

1. Pragmatic leader 5.58 0.47 24 2. Charismatic 5.32 0.47 23 1.91 3. Shared 5.22 0.47 22 2.62 ⁎ 0.71

Note. Bonferroni adjustment for multiple comparisons. ⁎ pb0.05.

677I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

Ashkanasy, N. M., & Tse, B. (2000). Transformational leadership as management of emotion: A conceptual review. In N. M. Ashkanasy, C. E. J. Hartel, & W. J. Zerbe (Eds.), Emotions in the workplace: Research, theory, and practice (pp. 221–235). Westport, CT: Quorum Books.

Barling, J., Weber, T., & Kelloway, E. K. (1996). Effects of transformational leadership training on attitudinal and financial outcomes: A field experiment. The Journal of Applied Psychology, 81(6), 827–832.

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182.

Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press. Borenstein, M., & Cohen, J. (1988). Statistical power analysis: A computer program. Hillsdale, NJ: Lawrence Erlbaum Associates. Brett, J. M., Shapiro, D. L., & Lytle, A. L. (1998). Breaking the bonds of reciprocity in negotiations. Academy of Management Journal, 41(4), 410–424. Burke, C. S., Stagl, K. C., Klein, C., Goodwin, G. F., Eduardo, S., & Halpin, S. M. (2006). What type of leadership behaviors are functional in teams? A meta-analysis.

Leadership Quarterly, 17, 288–307. Chartrant, T. L., & Bargh, J. A. (1999). The chameleon effect: The perception–behavior link and social interaction. Journal of Personality and Social Psychology, 76,

893–910. Cohen, S. G., Ledford, G. E., Jr., & Spreitzer, G. M. (1996). A predictive model of self-managing work team effectiveness. Human Relations, 49, 643–676. Connelly, S., Gaddis, B., & Helton-Fauth, W. (2002). A closer look at the role of emotions in transformational and charismatic leadership. In B. Avolio & F. J.

Yammarino (Eds.), Transformational and charismatic leadership: The road ahead. New york, NY: Elsevier Science. De Dreu, C. K. W., & Weingart, L. R. (2003). Task versus relationship conflict, team performance, and team member satisfaction: A meta-analysis. The Journal of

Applied Psychology, 88, 741–749. Dooley, R. S., & Fryxell, G. E. (1999). Attaining decision quality and commitment from dissent: The moderating effects of loyalty and competence in strategic

decision-making teams. Academy of Management Journal, 42(4), 389–402. Eagly, A. H., & Johnson, B. T. (1990). Gender and leadership style: A meta-analysis. Psychological Bulletin, 233–256. Eisenhardt, K. M. (1999). Strategy as strategic decision making. Sloan Management Review, 40(3), 65–72. Eisenhardt, K. M., Kahwajy, J. L., & Bourgeous, L. J. (1997). Conflict and strategic choice: How top management teams disagree. California Management Review, 39

(2), 42–62. Gardner, W. L., & Avolio, B. J. (1998). The charismatic relationship: A dramaturgical perspective. Academy of Management Review, 23, 32–58. Georgopolous, B. S. (1986). Organizational structure, problem-solving, and effectiveness. San Francisco: Jossey-Bass. Guth, W. D., & MacMillan, I. C. (1986). Strategy implementation versus middle management self-interest. Strategic Management Journal, 7, 313–327. Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1992). Primitive emotional contagion. In M. S. Clark (Ed.), Emotion and social behavior: Review of personality and social

psychology, Vol. 14. (pp. 151–177). Thousand Oaks, CA: Sage Publications. Hitt, M. A., & Tyler, B. B. (1991). Strategic decision models: Integrating different perspectives. Strategic Management Journal, 12, 327–351. House, R. J. (1998). Appendix: Measures and assessment for the charismatic leadership approach: Scales, latent constructs, loadings, Cronbach alphas, and

interclass correlations. In F. Dansereau & F.J. Yammarion (Eds.), Leadership: The multiple-level approaches. Stamford, Connecticut: JAI Press. House, R. J., & Shamir, B. (1993). Toward the integration of transformational, charismatic, and visionary theories. In M. M. Chemers & R. Ayman (Eds.), Leadership

theory and research. New York, N.Y.: Academic Press. Howell, J. M., & Frost, P. J. (1989). A laboratory study of charismatic leadership. Organizational Behavior and Human Decision Processes, 43(2), 243–269. Janis, I. L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes. Boston, MA: Houghton Mifflin. Janssen, O., Van de Vliert, E., & Veenstra, C. (1999). How task and person conflict shape the role of positive interdependence in management teams. Journal of

Management, 25(2), 117–142. Javidan, M., Bemmels, B., Devine, K. S., & Dastmalchian, A. (1995). Superior and subordinate gender and the acceptance of superiors as role models. Human

Relations, 48, 1271–1284. Jehn, K. (1994). Enhancing effectiveness: An investigation of advantages and disadvantages of value-based intragroup conflict. International Journal of Conflict

Management, 5, 223–238. Jehn, K. A. (1995). A multimethod examination of the benefits and detriments of intragroup conflict. Administrative Science Quarterly, 40, 256–282. Jehn, K. A., Chadwick, C., & Thatscher, S. M. B. (1997). To agree or not to agree: The effects of value congruence, individual demographic dissimilarity, and conflict on

workgroup outcomes. The International Journal of Conflict Management, 8(4), 287–305. Jehn, K. A., & Mannix, E. A. (2001). The dynamic nature of conflict: A longitudinal study of intragroup conflict and group performance. Academy of Management

Journal, 44(2), 238–251. Johnson, S. K. (2008). I second that emotion: Effects of emotional contagion and affect at work on leader and follower outcomes. Leadership Quarterly, 19, 1–19. Kirk, R. E. (1995). Experimental design: Procedures for the behavioral sciences. Toronto: Brooks/Cole Publishing Company, ITP. Kirkpatrick, S. A., & Locke, E. A. (1996). Direct and indirect effects of three core charismatic leadership components on performance and attitudes. The Journal of

Applied Psychology, 81, 36–51. Korsgaard, M. A., Schweiger, D. M., & Sapienza, H. J. (1995). Building commitment, attachment, and trust in strategic decision-making teams: The role of procedural

justice. Academy of Management Journal, 38(1), 60–84. Kotlyar, I., & Karakowsky, L. (2006). Leading conflict?: Exploring potential linkages between leader behaviors and group conflict. Small Group Research, 37,

377–403. Langfred, C. W. (2007). The downside of self-management: A longitudinal study of the effects of conflict on trust, autonomy, and task interdependence in self-

managing teams. Academy of Management Journal, 50, 885–900. Lazarus, R. S. (1993). From psychological stress to the emotions: A history of changing outlooks. Annual Review of Psychology, 44, 1–21. Maier, N. R. F. (1963). Problem-solving discussions and conferences: Leadership methods and skills. New York: McGraw-Hill. Manz, C. C., & Sims, H. P. J. (1987). Leading workers to lead themselves: The external leadership of self-managing work teams. Administrative Science Quarterly, 32,

106–128. Medina, F. J., Munduate, L., Dorado, M. A., Martínez, I., & Guerra, J. M. (2005). Types of intragroup conflict and affective reactions. Journal of Managerial Psychology,

20, 219–231. Mooney, A. C., Holahan, P. J., & Amason, A. C. (2007). Don't take it personally: Exploring cognitive conflict as a mediator of affective conflict. Journal of Management

Studies, 44(5), 733–758. Morgeson, F. P. (2005). The external relationship of self-managing teams: Intervening in the context of novel and disruptive events. The Journal of Applied

Psychology, 90, 497–508. Mumford, M. D. (2006). Pathways to outstanding leadership: A comparative analysis of charismatic, ideological and charismatic leaders. MAhwah, NJ: Erlbaum. Mumford, M. D., Antes, A. J., Caughron, J. J., & Freidrich, T. L. (2008). Charismatic, ideological and pragmatic leadership: Multi-level influences on emergence and

performance. The Leadership Quarterly, 19, 144–160. Peterson, R. S. (1997). A directive leadership style in group decision making can be both virtue and vice: Evidence from elite and experimental groups. Journal of

Personality and Social Psychology, 72(5), 1107–1121. Pfeffer, J. (1992). Managing with power: Politics and influence in organizations. Boston, MA: Harvard Business School Press. Phillips, J. S., & Lord, R. G. (1982). Schematic information processing and perceptions of leadership in problem-solving groups. The Journal of Applied Psychology,

67(4), 486–492. Priem, R. L., Harrison, D. A., & Muir, N. K. (1995). Structured conflict and consensus outcomes in group decision making. Journal of Management, 21(4), 691–710. Pruitt, D. G., & Rubin, J. Z. (1986). Social conflict: Escalation, stalemate, and settlement. New York: Random House. Riggio, R. E. (1986). Assessment of basic social skills. Journal of Personality and Social Psychology, 51(3), 649–660. Schwarz, N., & Clore, G. L. (1983). Mood, misattribution, and judgments of well-being: Informative and directive functions of affective state. Journal of Personality

and Social Psychology, 31, 513–523.

678 I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679

Schweiger, D. M., Sandberg, W. R., & Ragan, J. W. (1986). Group approaches for improving strategic decision making: A comparative analysis of dialectical inquiry, devil's advocacy, and consensus. Academy of Management Journal, 29(1), 51–71.

Schweiger, D. M., Sandberg, W. R., & Rechner, P. L. (1989). Experimental effects of dialectical inquiry, devil's advocacy, and consensus approaches to strategic decision making. Academy of Management Journal, 32, 745–772.

Schwenk, C., & Cosier, R. (1993). Effects of consensus and devil's advocacy on strategic decision-making. Journal of Applied Social Psychology, 23, 126–139. Sessa, V. I. (1996). Using perspective taking to manage conflict and affect in teams. The Journal of Applied Behavioral Science, 32(1), 101–115. Shamir, B., House, R. J., & Arthur, M. B. (1993). The motivational effects of charismatic leader: A self-concept-based theory. Organizational Science, 4(4), 577–595. Shamir, B., Zakay, E., Breinin, E., & Popper, M. (2000). Leadership and social identification in military units: Direct and indirect relationships. Journal of Applied Social

Psychology, 30, 612–640. Simons, T. L., & Peterson, R. S. (2000). Task conflict and relationship conflict in top management teams: The pivotal role of intragroup trust. The Journal of Applied

Psychology, 85(1), 102–111. Sy, T., Cote, S., & Saavedra, R. (2005). The contagious leader: Impact of the leader's mood on the mood of group members, group affective tone, and group processes.

The Journal of Applied Psychology, 90, 295–305. Tekleab, A. G., Quigley, N. R., & Tesluk, P. E. (2009). A longitudinal study of team conflict, conflict management, cohesion and team effectiveness. Group &

Organization Management, 34, 170–188. Thomas, D. C. (1999). Cultural diversity and work group effectiveness: An experimental study. Journal of Cross-Cultural Psychology, 30(2), 242–263. Tjosvold, D. (1993). Learning to manage conflict. New York: Lexington Books. Wageman, R. (1997). Critical success factors for creating superb self-managing teams. Organizational Dynamics, 26(1), 49–61. Wall, J. A. J., & Callister, R. R. (1995). Conflict and its management. Journal of Management, 21(3), 515–558.

679I. Kotlyar et al. / The Leadership Quarterly 22 (2011) 666–679