aritcles

The Critical Role of Conflict Resolution in Teams: A Close Look at the Links Between Conflict Type, Conflict Management Strategies, and Team

Outcomes

Kristin J. Behfar University of California, Irvine

Randall S. Peterson London Business School

Elizabeth A. Mannix and William M. K. Trochim Cornell University

This article explores the linkages between strategies for managing different types of conflict and group performance and satisfaction. Results from a qualitative study of 57 autonomous teams suggest that groups that improve or maintain top performance over time share 3 conflict resolution tendencies: (a) focusing on the content of interpersonal interactions rather than delivery style, (b) explicitly discussing reasons behind any decisions reached in accepting and distributing work assignments, and (c) assigning work to members who have the relevant task expertise rather than assigning by other common means such as volunteering, default, or convenience. The authors’ results also suggest that teams that are successful over time are likely to be both proactive in anticipating the need for conflict resolution and pluralistic in developing conflict resolution strategies that apply to all group members.

Keywords: team conflict, conflict management, team viability

A great deal of empirical and theoretical attention has been focused in the past decade on intragroup conflict and its impact on team performance (see De Dreu & Weingart, 2003, and Jehn & Bendersky, 2003, for reviews). Two recent developments in this line of research are the impetus for this article. First, the intragroup conflict literature has begun to question the notion that measuring levels of conflict by using the tri-partite classification (i.e., task, relationship, and process conflict) will reliably explain group per- formance. For example, a recent meta-analysis has cast doubt on the usefulness of this classification by questioning the widely theorized benefits of task conflict and suggesting that task conflict predicts similar negative outcomes to relationship conflict (De Dreu & Weingart, 2003). A number of scholars have also reported negative and contradictory associations between process conflict and performance (e.g., Jehn & Mannix, 2001).

The second and related development in the group conflict liter- ature is the move away from the notion that different types of conflict have a direct impact on performance—instead scholars are increasingly suggesting that various aspects of group process and group dynamics serve to ameliorate or exacerbate the impact conflict has on group outcomes generally (see Jehn, 1997, and Jehn & Bendersky, 2003, for a review). For example, task conflict

in the presence of trust is more likely to result in positive effects than where trust is low (Simons & Peterson, 2000). Building on these developments, we take a closer look at the effects of conflict resolution (i.e., vs. absolute level of conflict). We want to know how conflict resolution tactics are associated with the impact of task, relationship, and process conflicts on group outputs, and how different approaches to managing each type of conflict are asso- ciated with increases versus decreases in team performance. In doing so, we suggest that the way a team manages its conflicts is critically important for predicting team viability and over time performance.

The Role of Conflict Management in Predicting Team Outcomes

A number of scholars have argued that conflict management, and particularly conflict resolution, is an important predictor of the group and/or dyadic conflict–performance relationship (Jehn & Bendersky, 2003; Marks, Mathieu, & Zaccaro, 2001; Mathieu & Schulze, 2006; Tinsley, 2001; Weingart, 1992; Williams & O’Reilly, 1988). In theorizing about this key role for conflict management, previous research suggests that a process for man- aging conflict can help to reduce the negative impact of all types of conflict by restoring fairness, process effectiveness, resource efficiency, working relationships, and/or satisfaction of parties (e.g., Thomas, 1992). For example, Tjosvold (1991) has argued that a cooperative approach to conflict resolution allows conflict of a variety of types to be resolved in a way that is beneficial to the group. The procedural justice literature similarly supports the notion that a process of conflict management that allows groups to resolve their conflicts fairly will result in conflict of all types being

Kristin J. Behfar, The Paul Merage School of Business, University of California, Irvine; Randall S. Peterson, Organisational Behaviour, London Business School, London, United Kingdom; Elizabeth A. Mannix, The John- son Graduate School of Management, Cornell University; William M. K. Trochim, Policy Analysis and Management, Cornell University.

Correspondence concerning this article should be addressed to Kristin J. Behfar, The Paul Merage School of Business, The University of California, Irvine, Irvine, CA 92698. E-mail: kbehfar@uci.edu

CORRECTED JANUARY 30, 2008; SEE LAST PAGE Journal of Applied Psychology Copyright 2008 by the American Psychological Association 2008, Vol. 93, No. 1, 170 –188 0021-9010/08/$12.00 DOI: 10.1037/0021-9010.93.1.170

170

more effectively resolved, leading to more desirable group out- comes. The group value model, for example, posits that individuals benchmark their status in a group on the basis of how procedures are applied to them during group process (Lind & Earley, 1992). Team conflicts, whether latent or overt, often manifest themselves as process-based conflicts in the form of passive-aggressive be- haviors such as power plays, blaming, tardiness, or withholding information (Edelmann, 1993; J. Wall & Callister, 1995). The manner in which members treat each other while working through problems changes expectations for the next group interaction, for member satisfaction with the group (Lind & Tyler, 1988), and for member willingness to continue contributing to the group proac- tively (Jehn, 1997).

Moreover, there is also a growing body of evidence to suggest that, over time, the actions and reactions groups have to past performance and evolving group dynamics leave teams prone to some degree of chronic conflict, conflict spirals, and/or rigidity in their process management (Ancona & Chong, 1996; Arrow, Poole, Henry, Wheelan, & Moreland, 2004; De Dreu & VanVianen, 2001; Kuhn & Poole, 2000; Peterson & Behfar, 2003; Rahim, 2002). One critical key to understanding group performance lies in better understanding the choices groups make in choosing conflict resolution strategies to deal with this conflict. These choices are key to understanding a team’s ability to successfully learn and adapt task strategies to meet performance criteria (e.g., Argyris, 1982; Edmondson, 1999). If teams make conflict resolution choices that do not allow the team to resolve conflict effectively, those teams are likely to be prone to continuous, escalating con- flicts as members spend time reacting to provocative conflict behaviors of other team members rather than focusing on the task at hand. Teams that can adapt their processes appropriately, how- ever, are more likely to create teams where members are satisfied and performance is enhanced.

The conflict management process encompasses a wide range of activities, including communication, problem solving, dealing with emotion, and understanding positions (Brett, 2001; Pondy, 1992; Putnam & Poole, 1987). Conflict management behaviors have primarily been studied either as “individual styles” that are stable traits of individuals, as types of behaviors (such as threats or compromises), or as generalized behavioral orientations (e.g., avoiding, accommodating, compromising, competing, problem solving; see Folger, Poole, & Stutman, 2001, for a review of the multiple frameworks). Previous research in this tradition has shown that different conflict management orientations (e.g., col- laborating, competing, accommodating) affect success at the indi- vidual and team levels (Blake & Mouton, 1964; De Dreu, 1997; Morrill & Thomas, 1992; Putnam & Poole, 1987; Rahim, 1983; Ruble & Thomas, 1976). For example, cooperative or collabora- tive orientations increase effectiveness in managing task conflict (De Dreu, 2006; Tjosvold, Hui, & Yu, 2003). Relationship con- flict, however, is typically described as needing to be managed differently, with some studies recommending avoidance of rela- tionship conflict altogether (e.g., De Dreu & VanVianen, 2001) and other studies reporting that avoidance orientations increase negative emotion (Desivilya & Yagil, 2005). The present article builds on this tradition but takes a more fine-grained look at conflict management tactics (as opposed to generalized strategies such as “cooperation”) by trying to get a better understanding of the effects of specific conflict resolution strategies on group out-

comes. We ask questions such as how do some teams manage relationship conflict and avoid its usual negative impact, while others do not? More specifically, are there particular conflict resolution tactics that allow for the effective resolution of relation- ship conflict? Why does task conflict manifest itself as construc- tive debate in some teams but as open fighting in other teams? In other words, are there specific conflict resolution tactics that a team can chose that will allow for effective management of task conflict? Do certain conflict resolution tactics work across all types of conflict? Or does each type of conflict require different conflict resolution strategies?

This Study: Linking Conflict Resolution Strategies With Team Performance and Member Satisfaction

Our purpose in this study is to examine specific conflict reso- lution strategies in groups to better understand their potential effects on group outcomes (e.g., performance and satisfaction). To accomplish this, we look at teams with consistently high or in- creasing group performance and member satisfaction and consis- tently low or decreasing group performance and member satisfac- tion over time in order to see which conflict resolution strategies are associated with each of three commonly measured conflict types: task, relationship, and process conflict. Task or cognitive conflict is disagreement over differences in ideas, viewpoints, and opinions pertaining to the group’s task (Amason & Sapienza, 1997). Relationship conflict is disagreement resulting from inter- personal incompatibilities, which includes affective components such as feeling tension and friction. Process conflict is conflict about dividing and delegating responsibility and deciding how to get work done (Jehn, 1997, p. 540). Each of these conflicts has been theorized to result in different group dynamics, but very little is known about specific strategies teams employ to manage the different types of conflict or the efficacy of these strategies. In this study, we are specifically interested in understanding (a) how groups perceive that they have managed each type of conflict and (b) how different conflict management approaches are associated with strong or weak outcomes in team performance and satisfac- tion. By understanding how groups experience and respond to different types of conflicts, we also hope to understand how performance spirals (both upward and downward) are established (cf. Lindsley, Brass, & Thomas, 1995).

Because relatively little research has examined the impact of intragroup (i.e., as opposed to individual member) conflict reso- lution strategies, we begin by examining the strategies that auton- omous work groups create to self-manage their team conflict. We chose to investigate conflict management in autonomous teams for two reasons. First, autonomous groups have become more preva- lent in the past 20 years, although often called by different names, including self-managing work teams, leaderless groups, high per- formance teams, and shared leadership teams (Cohen & Ledford, 1994). Autonomous teams are also widely used in academic set- tings both as classroom tools as well as for research samples in academic research (Loyd, Thompson, & Kern, 2005). Second, in theory, autonomous teams are particularly well suited to illustrate how the management of team processes, such as communication and conflict management, can enhance versus detract from team viability. Because decision-making power about team processes is shifted from a manager directly to team members, autonomous

171CONFLICT MANAGEMENT IN TEAMS

team members are also responsible for managing the communica- tion and conflict that results from task decisions and subsequent group processes. Since team viability, or sustainable success over time, depends on a team’s ability to adapt in response to changes in the environment, to learn from feedback, and to resolve conflicts when faced with differences between old ways of doing things and new performance expectations (Arrow, McGrath, & Berdahl, 2000; Hackman & Morris, 1975; Ilgen, John, & Jundt, 2005; McGrath, 1991; Okhuysen & Eisenhardt, 2002; Poole, Siebold, & McPhee, 1996; Vancouver, 2000), autonomous teams should be particularly well placed to display this type of adaptability. In practice, however, the empirical evidence has been mixed (see Beekun, 1989; Cohen & Ledford, 1994; Cummings & Griggs, 1977; Pasmore, Francis, & Haldeman, 1982; T. Wall, Kemp, Jackson, & Clegg, 1986). Studies have found that autonomous teams exhibit productivity declines over time, higher turnover rates, coordination problems, and declining member motivation (Cordery, Mueller, & Smith, 1991; Goodman, Devadas, & Hugh- son, 1988; Guzzo, 1982; Howell, Bowen, Dorfman, & Podskaoff, 1990; Moorhead, Neck, & West, 1998; Steiner, 1972). Therefore, autonomous teams represent a critical test of how the use of specific conflict resolution tactics either facilitates or hinders ef- fective conflict resolution.

Method

Research Setting and Participants

The research sample was the entire 1st-year MBA class of 252 students, or 65 study teams, at an East Coast graduate school of management. The participants worked in the same teams across all four of their first semester core curriculum classes, and the team portion of their work counted for at least 40% of each individual’s grade in each class. This sample was chosen because it constituted newly forming, intact autonomous groups with no previous his- tory. Their workload was sufficiently heavy to make task interde- pendence necessary, and their performance outcomes (i.e., grades) were important to individuals’ academic standing (i.e., there were probation consequences for low team grades) and job prospects (for both recruiting and tuition reimbursement). In addition, the school is relatively small, so it was highly likely that the team members would work together again in the course of their MBA program. Thus, while these were student teams, the atmosphere and consequences of the work were a reasonable simulation of autonomous business teams with both task-related work as well as social relationship and reputation consequences if the groups failed. The fact that the teams were newly forming and began their work with the same baseline resources was also important to this study in terms of differentiating the effectiveness of team conflict management strategies.

The students were an average age of 29 years; 27% women, 73% men; 34% born outside the United States; with an average of 4.7 years of full-time work experience after completing undergrad- uate degrees in a wide range of managerial and technical positions. Groups were randomly assigned with 3– 4 members, with the caveat that each team contained at least 1 student born outside of the United States. With the exception of a few students, most of the foreign-born students had spent between 4 and 10 years living, attending university, and/or working in the United States. Teams

did not have formally appointed leaders, and they were jointly responsible for the outcomes of the group. None of these demo- graphic variables were significantly associated with our key per- formance measures and so were not analyzed further.

Although students used their same study groups to do problem sets in accounting and economics, the measurement used for performance in this study was from their graded group assign- ments in the organizational behavior core course because problem sets had little variance in performance (i.e., they were mostly correct). The organizational behavior course assignments consisted of two five-page case analyses, worth 20% and 30%, respectively, of their final course grades. The two case analysis assignments represent non-routine tasks, as teams were responsible for identi- fying the management problem represented in the case and for recommending management actions for resolving the problem on the basis of research and theory presented in the core organiza- tional behavior course. The core course met three times per week over 10 weeks. The first paper was due in Week 4, the second in Week 9. Students were given cases to prepare and were asked to “demonstrate your ability to apply what you have learned in selective and creative ways to analyze and diagnose organizational problems and solve them.”

Procedure and Measures

Overview of Procedure

The data analyzed below are a mix of qualitative and quantita- tive analyses. We specifically make links between the qualitative process and quantitative outcome data. Open-ended survey ques- tions were used to gather the qualitative data about team norms, conflicts, and conflict management, which were analyzed both by study participants and academic experts. We chose the open-ended survey format because it elicits brief yet detailed team member accounts, minimizes demand characteristics because it does not constrain team members’ options to discuss certain types of con- flict/management behaviors, and usually captures the most salient or important aspects of team processes (Fine & Elsbach, 2000; Miles & Huberman, 1994). We chose not to use existing conflict management inventories in this study because we were interested in capturing detail about team-level development and application of different strategies to different types of conflict. Therefore, we rely on a more participant-driven method of inquiry. Survey data were also collected to assess satisfaction.

Measures

Surveys. The teams were administered two surveys that in- cluded both open-ended questions and close-ended measures. The first survey was administered after teams had completed their first group assignment but before they received a grade for that assign- ment (Time 1). Measurement was taken prior to the participants receiving their grades to ensure that they reported their observa- tions about their group experiences instead of their reactions to the group’s grade. The Time 1 survey contained an open-ended ques- tion about team norms and a close-ended measure of satisfaction, along with other items relevant to the course but not used in this study. The open-ended question asked respondents to describe their team norms, defined as written or unwritten patterns of

172 BEHFAR, PETERSON, MANNIX, AND TROCHIM

beliefs, attitudes, communication, and behaviors that become es- tablished among team members (cf. Feldman, 1984; Guzzo & Shea, 1992). While the focus of the study was on conflict resolu- tion, we wanted to cross-reference normative data with our reso- lution findings as a way of validating our conclusions and to compare early versus later team norms as a reflection of how effective the strategies were (in addition to performance and sat- isfaction scores; Jick, 1979). The teams received their grade for their first group assignment only after Time 1 surveys were col- lected.

The Time 2 survey was identical to the Time 1 survey but also included two open-ended questions asking team members to de- scribe the conflicts (if any) they had since their first group assign- ment and how they addressed or managed those conflicts. During the term, respondents were directed to think about their team experience since their first assignment (T1 measurement) and describe “What types of conflicts or disagreements arose in your team?” and “How did your team manage or resolve these conflicts or disagreements?” The Time 2 survey was administered after they completed their second group assignment but before they received a grade for that assignment or for their final course grade. We chose to collect data about conflict and conflict management only at Time 2 because we were interested in capturing conflict man- agement strategies during a complete team performance episode (Marks et al., 2001), defined as distinguishable periods of time over which performance accrues and feedback is available (p. 359). Since teams are not particularly aware of the effectiveness of their process/task strategies until after they receive feedback (Hackman, 1990), and the impact of performance feedback is often that of increased conflict (e.g., Peterson & Behfar, 2003), we were interested in examining conflict resolution strategies when conflict experiences were particularly salient and as teams learned how to adapt or transition their task strategies for their next assignment (Marks et al., 2001). In contrast, team norms are a relatively stable team property (Bettenhausen & Murnighan, 1985), so we mea- sured them at both time periods as a cross-check to the effective- ness of conflict management strategies. We recognize that concen- trating on one performance episode limits the conclusions we can draw about conflict management and team viability over time. However, we did so in order to uncover how conflict management is associated with the beginnings of positive versus negative out- come trends.

At Time 1, 244 responses (8 non-responses) were collected, for a response rate of 96.8%; the Time 2 data included 225 responses (27 non-responses), for a response rate of 89.3%. Teams with less than two-thirds of the members answering the surveys across both time periods were dropped from the study (e.g., a 4-person team

needed at least 3 respondents to be included). There was no apparent pattern in the non-responses—members of these teams came from a wide range of performance and satisfaction scores. Of the teams, 57 responded at the required level on both surveys to be included in our dataset here, and 8 did not respond and were subsequently dropped from all analyses.

Team outcome measures. In order to investigate associations with team outcomes, team grade and team member satisfaction were treated as “outcome” measures. Because conflict manage- ment represents a convergence of cognitive and emotional forces on group decision making (Folger et al., 2001; Hackman & Morris, 1975; J. Wall & Callister, 1995), team grade was the variable we chose to represent a group’s ability to successfully adapt its task management strategies. Satisfaction was the variable we chose to represent how well a group was managing the social or affective side of its work.

Grades with a possible range of 60 to 100 were assigned by one of two class professors who taught identical content but different sections of the course. Both professors used the same grading criteria and point systems for grading the cases. The performance scores at Time 1 ranged between 70 and 95, with a standardized mean and median of 80 for both professors. The performance scores at Time 2 ranged between 70 and 95, with a standardized mean and median of 84 for both professors. See Table 1 for descriptive statistics on these data.

Satisfaction was measured with five items rated on a 9-point Likert scale adapted from Peterson (1997) that included asking participants how satisfied they were working with the team, how much they liked other team members, to what extent the other people on the team were generally friendly, if they would like to work with their team again in the future, and how satisfied they thought their fellow teammates were with being a member of the team. Cronbach’s alpha yielded a reliability coefficient of .91 at Time 1 and .92 at Time 2 for the satisfaction measure. All scores were aggregated to the team level by calculating the team mean score. See Table 1 for descriptive statistics on these data.

To assess the appropriateness of aggregating the satisfaction items to the team level, the within-group agreement index, rwg (James, Demaree, & Wolf, 1984), and intraclass correlation coef- ficient, ICC(1), were calculated for Time 1 and Time 2. The rwg index ranges from 0 to 1, where values closer to 1 mean greater within-group agreement. Klein et al. (2000) and numerous others recommend a value of rwg � .70 to justify aggregating to the group level, although the .70 cutoff is only a “rule of thumb” (Lance, Butts, & Michels, 2006). Values exceeded this cutoff: at Time 1, rwg� .94; and at Time 2, rwg � .92. ICC(1) is also used to determine if aggregation is warranted and typically results in much

Table 1 Descriptive Statistics for Group Outcome Measures

Variable Time 1

performance grade Time 2

performance grade Time 1

satisfaction Time 2

satisfaction

Mean 80 84 7.6 7.0 Median 75 85 8.0 7.2 Standard deviation 7.3 7.5 .98 1.3 Range (min/max) 25 (70/95) 25 (70/95) 4.7 (4.1/8.8) 6.6 (2.4/9)

Note. Performance was measured by students’ grades.

173CONFLICT MANAGEMENT IN TEAMS

lower values (a value � .12 is generally considered good; James, 1982). Values exceeded this cutoff: Time 1 � .41; Time 2 � .47. The above findings indicate that there was more significant agree- ment within groups than between groups, and, therefore, the ag- gregated satisfaction score was deemed appropriate.

Qualitative Data Analysis

The purpose of analyzing the qualitative responses was to ex- plore three questions: (a) What types of conflict resolution strat- egies did the teams use to address different types of conflict (task, relationship, and process)? (b) How are those strategies associated with positive versus negative outcome trends between Time 1 and Time 2? and (c) Was the success/failure of conflict management reflected in changes in team norms? The qualitative data from the conflict-type open-ended question were coded by academic ex- perts. The data from the conflict management question were ana- lyzed by participants. The responses to the team norms questions were coded by two professional managers of teams, each with 10� years of team management experience (more detail in each section below).

Conflict and Conflict Management

The responses to both of the qualitative questions about conflict and conflict management were typically one to three sentences containing one or two ideas about conflict and conflict resolution. All respondents wrote something in for this question, even if it was to indicate that the team experienced very little conflict or did very little to manage it. Therefore, for each team we were able to separate the responses into a list of statements about conflict and a list of resolution strategies employed. The analysis of these data consisted of three stages: (a) an expert rating of the extent to which each conflict statement (responses to the first open-ended ques- tion) was related to task, relationship, and process conflict (Jehn, 1997); (b) a categorical coding by participants of conflict resolu- tion strategies (responses to the second open-ended question) with the concept mapping methodology (explained below; Jackson & Trochim, 2002); and (c) examination of how different conflicts and the management strategies employed to resolve those conflicts were associated with different team outcome patterns. Thus, for each respondent, there was a conflict statement, an expert coding of the type of conflict mentioned in that statement, a conflict resolution statement, and a categorical coding of that resolution statement done by the participants from the concept mapping analysis—all of which were linked to team-level trends for per- formance and satisfaction.

Classifying Conflict Type by Using Expert Ratings

All of the conflict responses from the first open-ended question were decomposed into single statements by Kristin J. Behfar. A sample of the 50 statements was also decomposed by a hypothesis- blind coder with strong agreement (i.e., Cohen’s � � .72), sug- gesting acceptable reliability in identifying discreet units of anal- ysis. For example, 1 participant’s response was “Commitments were made initially but were not followed up or backed up with the required efforts. Opinions strictly related to projects and write-ups were often interpreted by my teammates as personal comments.”

This response was decomposed into two separate statements: (a) Commitments were made initially but were not followed up or backed up with the required efforts, and (b) opinions strictly related to projects and write-ups were often interpreted by my teammates as personal comments. Next, a group of 24 academic experts in groups and teams research completed a 9-point Likert scale rating for a subset of the 235 conflict statements generated according to how closely they were related to a specific type of conflict (e.g., task, relationship, or process; Hinkin & Tracey, 1999). These experts were faculty members and doctoral students in management, social psychology, and industrial labor and rela- tions departments. Each statement was evaluated by at least 2 experts on a rating scale from 1 (not at all related) to 9 (completely related) as to how related it was to each type of conflict. They were given definitions of each type of conflict as presented in Jehn (1997). The interrater reliability for the expert raters was assessed with Cronbach’s alpha, yielding reliable ratings of � � .84 for task conflict, � � .79 for relationship conflict, and � � .75 for process conflict.

Classifying Conflict Management by Using Participant Concept Mapping

The concept mapping method can best be thought of as “par- ticipatory content analysis” (Jackson & Trochim, 2002) and as a hybrid between traditional content analysis and semantic mapping analysis. We chose this method to analyze the conflict resolution responses because the research objective was to understand how participants themselves (rather than the researchers) think about conflict resolution strategies. Concept mapping as applied to qual- itative data analysis (Jackson & Trochim, 2002) combines explor- atory statistical analysis with participants’ judgments to produce clusters of similar thematic categories by using multidimensional scaling and cluster analysis. The analysis is a five-step process: (a) determining units of analysis, (b) participant sorting of units, (c) multidimensional scaling analysis, (d) cluster analysis, and (e) cluster labeling. Beyond creating units of analysis from the initial responses, the researchers did not make any coding or analysis decisions.

Determining units of analysis. Units of analysis were created from the statements generated by respondents in response to the second open-ended survey question, “How did you resolve those conflicts?” Each raw response from a respondent was typically one to two sentences long. Units of analysis were created by separating respondents’ answers into single statements, each containing only one idea about conflict. For example, 1 respondent’s answer to the conflict resolution was “The issue was discussed and everyone was asked an opinion. Basically it was the way we approached the case assignments. [and] We had two people work on the first assign- ment and made turns on the next one.” This response was broken into two separate statements: (a) The issue was discussed and everyone was asked an opinion. Basically it was the way we approached the case assignments; and (b) we had two people work on the first assignment and made turns on the next one. All of the conflict resolution responses from the open-ended question were decomposed into single statements by Kristin J. Behfar. This process resulted in 210 statements about conflict resolution—an average of three to four statements per team. These 210 statements were matched with their corresponding conflict statement for anal-

174 BEHFAR, PETERSON, MANNIX, AND TROCHIM

ysis by a respondent identification and team number. A sample of the 50 statements was also decomposed by a hypothesis-blind coder with strong agreement (i.e., Cohen’s � � .90), suggesting acceptable reliability in identifying discreet units of analysis.

Participant sorting of units of analysis. To avoid introducing researcher bias to the remaining steps of the concept mapping analysis, MBA students were used as decision makers. Second- year students were chosen instead of the original study respondents to protect privacy. Second-year students are a reasonable proxy for the original respondents because they are, as a group, demograph- ically virtually identical and they have experienced the same courses and have been members of similarly composed teams. We gave 15 students, 9 male and 6 female, who responded to a recruiting poster, a set of cards with statements (i.e., units of analysis) on them and instructed them to sort cards containing similar ideas together into piles (cf. Jackson & Trochim, 2002, on the method). The students worked individually, and there was no limit to the number of piles they could create. They were asked to give each of their piles a name. The only restriction was that they could not create a “Miscellaneous” pile—if they thought a state- ment did not belong with any of the others, they were instructed to leave it in its own pile.

Multidimensional scaling analysis. A multidimensional scal- ing analysis on the sorting was done to create a map of conceptual similarity between the statements that visually displayed the sim- ilarity judgments of the sorters. A 210 � 210 binary square matrix (rows and columns represent conflict resolution statements) was created for each individual sorter. Cell values represented whether or not a pair of statements was sorted by a particular coder into the same pile (i.e., yes vs. no; 1 vs. 0 coding). These individual matrices were then aggregated by adding together all 15 of the individual matrices. From the aggregated matrices, multidimen- sional scaling created coordinate estimates and a two-dimensional map of distances between the statements based on the aggregate sorts of the 12 coders. A two-dimensional solution was chosen because it provides the most useful foundation for a cluster anal- ysis (Kruskal & Wish, 1978).

Cluster analysis. Cluster analysis was conducted on the mul- tidimensional scaling coordinates. Two additional 2nd-year MBA students worked together to choose the cluster solution (i.e., num- ber of clusters) that they felt most accurately represented the structure of the data. They made their final decisions by looking at the cluster dendogram and discussing whether or not the contents of clusters merging at each solution were conceptually similar enough to merge. Their decisions about the final number of con- ceptual clusters to choose represents a final solution for the data.

Cluster labeling. After a final solution was chosen, the same two MBA students re-examined the statements in each cluster, as well as the names the original sorters had given each of their piles, to determine a label that best represented the content of the clusters. They then chose what to name or label each cluster. All cluster groupings and the language of the cluster labels chosen originated from the sorters’ labels or from the participants’ statements.

Summary Analysis: Linking Conflict Type, Conflict Management, and Outcome Patterns

To examine the link between conflict types, conflict manage- ment strategies, and positive versus negative outcome patterns

during the teams’ life cycle, we divided the qualitative data into four categories: (a) teams with consistently high or increasing performance and satisfaction; (b) teams with consistently high or increasing performance but consistently low or decreasing satis- faction; (c) teams with consistently low or decreasing performance but consistently high or increasing satisfaction; and (d) teams with consistently low or decreasing performance and satisfaction. To create these categories, we first assigned codes to each team on the basis of their outcomes relative to the mean (above or below) between the two time periods. For example, a team that was below the mean at Time 1 but above the mean at Time 2 was coded as increasing. Teams that were above the mean at Time 1 but below the mean at Time 2 were coded as decreasing. Teams that were above or below the mean at both time periods were coded as consistent. This was done for both grade and satisfaction scores at Time 1 and Time 2. We chose to classify teams by their relative position to the mean for two reasons.1 First, it provides a relatively objective cutoff point for assessing a meaningful change, or an upward versus downward pattern, in a dense outcome distribution. Second, it allows us to capture “high” versus “low” start and end points for each team during the performance episode we are investigating. The object of this analysis was to identify associa- tions between conflict management strategies and outcome pat- terns—not to identify characteristics of strategies associated with one single high or low performance outcome. We believe this is meaningful because teams performing well early need to remain vigilant about their process effectiveness and resist the temptation to blindly rely on previously successful tactics (e.g., Hackman & Wageman, 2005). Therefore, the lumping of increases with con- sistently high scores but decreases with consistently low scores represents strategies that allow for continuous growth versus con- tinuous decline.

Intrateam Agreement

In order to construct a team-level account about conflict and conflict management from the individual members’ accounts, we followed a two-step decision rule based on within-team concept frequencies. For conflict type, frequency counts were based on the expert rater’s classification of the conflict units, and for conflict management frequency counts were based on the concept mapping classification of the resolution units (Miles & Huberman, 1994). In order to minimize validity threats associated with frequency counts, we followed a two-step inclusion/exclusion decision method. First, because of the open-ended “free-recall” nature of the data and the way we created units of analysis, there is the potential for a particularly angry, strong-minded, or verbose team member to generate signifi- cantly more units than might those with more moderate tendencies (Geer, 1991). For example, 1 team with 3 survey respondents gener- ated four conflict units. Member 1 was associated with two of those units, both about personality conflicts. This has the potential to over- represent the distribution of meaning about certain types of conflicts in the sample. Therefore, we followed the convention of constraining frequency counts to only one unique concept per team member (Kraut, 1996). In the above example, this means that we would

1 The same analysis was done with change scores (Time 2 – Time 1) to create categories for comparison. The results were virtually identical.

175CONFLICT MANAGEMENT IN TEAMS

eliminate one of Member 1’s units about personality conflict. How- ever, if Member 1 had generated one statement about personality conflict and one about task conflict, both units would have been included at this stage. This resulted in the elimination of only six conflict units and five associated conflict resolution units (one of the eliminated conflict units did not have an associated resolution unit). While only six units did not meet inclusion criteria, their exclusion does help correct for potential overrepresentation threats and any unitizing artifacts.

The second step in excluding units was intended to roughly capture intragroup agreement about conflict experiences. Recall that we included teams in the analysis with at least two-thirds of members responding. To assess agreement about the types of conflict each team experienced, we examined the number of indi- viduals that mentioned different types of conflict. We followed a 50% rule for inclusion in 4-person teams and a two-thirds rule for inclusion in 3-person teams (if there were only 2 respondents from a 3-person team, both or 100% had to mention a type of conflict). This exclusion criterion resulted in the elimination of 30 conflict units and their associated conflict management units: 12 about process conflict, 8 about relationship conflict, and 10 about task conflict. After both steps, 174 of the original 210 conflict man- agement units (a reduction of 36 units) and their associated conflict statements were included in the summary analysis.

Interteam Agreement

In order to draw conclusions about how teams in each outcome pattern category addressed the different conflicts, at least one-third of the teams in each outcome pattern category had to report applying that strategy. The one-third criterion was set because there were three types of conflict coded for task, relationship, and process. The majority of teams reported addressing only one type of conflict, but 11 teams (19% of the sample) reported addressing two types of conflict.

Team Norms Data

The data from the team norms questions were used to lend support to the conclusions drawn from the resolution strategies analysis. To better place the strategies in evolving normative team dynamics, we compared the differences between Time 1 and Time 2 accounts of team norms. Two independent coders blind to the purpose of this study coded all of the team norm accounts for mention of trouble or dysfunction. The coders were one male from the engineering field and one female manager from the marketing field, each with over 10 years of team management experience. They were chosen on the basis of their 10� years of professional experience with both leading and being a member of more than 30 teams. They used a binary coding scheme: Teams received a 1 if trouble was mentioned and a 0 if no trouble was mentioned. Examples of trouble or dysfunction included explicit and specific mention of problems such as coalitions forming against a team member, underlying tension among members, ongoing process problems, and so on. Some examples are (a) “Sometimes we lost focus. Also near the end, we all were so busy, we probably did not try as hard to make things work—we just wanted to get the project over with!”; (b) “We had more problems in the second half of the course. The team meeting was not attended by everyone. Team

members walked out of the team when there was a disagreement. An optional case was not completed due to a team member’s different opinion. Had we had something in writing as the team norm, we’d be better off dealing with these issues”; and (c) “The norm is, and always has been, that it is not easy going for us. Our problems never really went away.” Examples of norms coded as having no mention of trouble include (a) “We have no problem, open, friendly team. Goal is to learn together. Kind, considerate, ‘things are under control’”; (b) “A strong intent to seek consensus. Open and frank communication. Willingness to listen to as well as provide views on all group work”; and (c) “Everyone took the team seriously. Nobody ever showed up late. A lot of respect towards the other team members. Great friends. Fun people, but the underlying key to the success was the mutual respect we held.”

The coding was done on both Time 1 and Time 2 norms. Cohen’s kappa was used to assess the interrater reliability, which met acceptable agreement of .81 at Time 1 and .96 at Time 2 (Fleiss, 1981). On the basis of the coding, we established four categories of norms patterns: (a) Teams mentioning trouble at both Time 1 and Time 2; (b) teams mentioning trouble at Time 1 but not Time 2; (c) teams not mentioning trouble at Time 1 but mentioning Trouble at Time 2; and (d) teams not mentioning trouble at Time 1 and Time 2. A chi-square test was performed to determine whether the frequency of patterns in team norms differed between the four categories of team outcome patterns discussed earlier (e.g., consistently high / increasing performance and satisfaction).

Results

The results section is organized as follows: We first discuss the results of the participant-based concept mapping analysis. We then report how different types of conflict (task, relationship, or pro- cess) mesh with choice of conflict resolution strategy and different outcome patterns. We then report how the team norms patterns link to the outcome patterns to verify our outcome patterns.

Concept Mapping of Conflict Resolution Strategies

The concept mapping analysis of conflict resolution statements resulted in seven categories of conflict resolution strategies: vot- ing, compromise or consensus, discuss or debate, open communi- cation, idiosyncratic solutions, avoided or ignored, and rotating responsibilities. The final map from the concept mapping analysis of conflict resolution strategies is presented in Figure 1.

The Concept Map

When interpreting the final map, note that each statement gen- erated by the respondents is represented as a point on the map that is included in a cluster. The position of each cluster on the map (e.g., top, bottom, right, left) is not meaningful— only the distance or spatial relationship between them is relevant. The proximity of the clusters to each other represents how similar the coders/sorters judged the statements within them to be. Clusters that are farther apart on the map contain statements that were sorted together less often than those that are closer together. The statements in the rotating responsibilities cluster, for example, were almost never sorted with those in the open communication cluster. However, the discuss or debate and open communication clusters are close

176 BEHFAR, PETERSON, MANNIX, AND TROCHIM

together. They both contain statements about team discussion, but the discuss or debate cluster focuses on substantive task debate (e.g., debating alternatives), and open communication focuses on the affective tone of the discussion (e.g., ego-invested vs. amicable task-centric debate). It is reasonable that the participants viewed these as closely related but conceptually distinct. The shape and size of a cluster generally represents whether it is a broad or narrow conceptual area but does not allow for meaningful inter- pretation (e.g., the size of a cluster does not represent the number of statements in a cluster).

Cluster Content

Representative statements from each cluster are displayed in Table 2. While the content of each cluster represents conceptual similarity (e.g., statements about voting in the voting cluster), there were differences within each cluster in how the different strategies were applied to different types of conflict. In the voting category, for example, there were ideas about using voting to address pro- cess problems versus resolving impasses on task debates. In the compromise or consensus cluster, ideas ranged from reaching a secure and collective compromise, to compromising too quickly to preserve relationships, to using compromise as a way to avoid discussion. In the discuss or debate cluster, ideas ranged from debating ideas with evidence-driven discussion to debating about controversial team process issues. The open communication clus- ter contained ideas about the affective tone of discussion, ranging from ego-invested to amicable. The idiosyncratic solutions cluster contained ideas about how teams punished or prevented conflicts, ranging from creating rules, to making threats or direct confron- tations, to empowering 1 person as a mediator. In the avoided/ ignored cluster, there were statements about strategies to avoid/ prevent conflict from escalating versus ignoring the existence of a conflict. The rotating responsibilities cluster contained ideas about how to manage conflicts by dividing work and team responsibil- ities, ranging from pre-structuring work assignments to structuring

in reaction to previously missed assignments. The next stage of the analysis identified how these differences were associated with increases and decreases in team outcomes.

Linking Conflict Type, Conflict Management Strategies, and Outcome Patterns

The results are reported below according to outcome patterns: (a) teams with consistently high or increasing performance and satisfac- tion; (b) teams with consistently low or decreasing performance and satisfaction; (c) teams with consistently high or increasing perfor- mance but consistently low or decreasing satisfaction; and (d) teams with consistently low or decreasing performance but consistently high or increasing satisfaction. Overall, 21 teams, or 37% of the sample, demonstrated consistently high or increasing outcome trends. These teams generated 55 statements about conflict resolution, which con- stituted 32% of the sample statements. Fourteen teams, or 25% of the sample, demonstrated consistently low or decreasing outcome trends. These teams generated 40 statements about conflict resolution, which constituted 23% of the sample statements. Eleven teams, or 19% of the sample, demonstrated consistently high or increasing performance but consistently low or decreasing satisfaction. These teams generated 46 statements about conflict resolution, which constituted 26% of the sample statements. Eleven teams, or 19% of the sample, demonstrated consistently low or decreasing performance but consistently high or increasing satisfaction. These teams generated 33 statements about conflict resolution, which constituted 19% of the sample statements.

Changes in Team Norms

We looked next to verify these outcome patterns with data col- lected on team norms. A significant chi square, �2(9, N � 57) � 30.91, p � .01, was found in the 4 � 4 matrix, revealing that there are statistically significant differences in the proportions of norm changes according to outcome pattern when compared with chance (Nunnally & Bernstein, 1994). Significance testing of these differences in pro- portions for each type of conflict was found by examining standard- ized residuals for values greater than or equal to 2 (Agresti, 1990). The standardized residual values are approximately normally distrib- uted, and a value of �2 or –2, at the � � .05 level, indicates that that cell of the row � column table contains a larger or smaller observed value than was expected (Agresti, 1990). Results are consistent with the conflict management findings. Table 3 illustrates that the consis- tently high/increasing grouping had a lower proportion of teams with troubled norms at Time 1 and Time 2. In contrast, the teams with consistently high/increasing performance but consistently low/ decreasing satisfaction and the teams with consistently low/decreasing performance and satisfaction had a higher proportion of team men- tioning troubled norms at Time 1 and Time 2. The norms statements for each group of teams reflect similar trends as the conflict manage- ment statements. For example, the teams with consistently high/ increasing performance and satisfaction described norms about pro- active problem solving, foreseeing and preventing problems, and learning to work with individual member’s unique traits. The consis- tently high/increasing performance but consistently low/decreasing satisfaction teams described norms about prevention, increased pro- cess structure, and role clarification or rules. The consistently low/ decreasing performance and satisfaction teams described norms of confusion, disorientation, and expectations for trouble. These results

Voting

Idiosyncratic Solutions

Compromise o Consensus

r

or Ignored Open Communication

Discuss or Debate

Avoided

Rotating Responsibilities

Figure 1. Categories of conflict resolution strategies. The position of the clusters on the map is based on a multidimensional scaling analysis: Only the distance between clusters is meaningful (not their physical location). The more proximal the clusters, the greater conceptual relation they have to each other. Size of the clusters does not indicate statement frequency. The cluster solution and cluster names were chosen by study participants.

177CONFLICT MANAGEMENT IN TEAMS

Table 2 Representative Statements From Each Cluster of Conflict Resolution Strategies on the Concept Map

Cluster name and description Representative conflict resolution statements

Voting “We tended to let people express their points of view, but typically majority ruled at decision time.” Ideas about using voting

procedures “If no one volunteered for an assignment we voted them in.” “We discussed opinions until frustrated and then voted.” “Only when we were out of time and at an impasse we voted.” “Only a few times it was required to informally get ‘votes’ for support to resolve conflict.” “We recognized when it was time to ‘agree to disagree’ and then voted.”

Compromise or consensus

Ideas about how groups reached agreement

“We talked each issue deeply and carefully ‘til we reach the consensus.” “We tried to compromise so as to not jeopardize our relationship.” “Compromise—mostly I just said “whatever” which probably wasn’t great team behavior.” “We reached consensus on who did what.” “We talked about our schedules and then compromised. One member gave up a weekend, another agreed to get up

early one morning.” “We recognized differences in work styles and tried to compromise our preferences.” “We would discuss in detail everyone’s views until we got a consensus (sometimes more than I could take.)”

Discuss or debate Ideas about discussing

and debating ideas/ opinions

“If one member thought one thing and another disagreed, we just discussed the pros/cons and quickly came to a conclusion.”

“We went through each solution to determine which was the easiest to defend with the information in the case.” “We met and voiced our concerns regarding the quality and effort put into the paper.” “Making sure everyone had a chance to fully explain their views and then went through a logical process of

discussion to pick the most convincing argument.” “Although the majority of time was spent on determining how to convince one or more members to the ideas.”

Open communication Ideas about the level of

affect or emotion in team discussions

“No one was hard set on making their point, everyone was open.” “The team members took the arguments positively and not personally.” “To resolve conflict we tried to keep open-minded and learn from one another.” “Discussions were very amicable.” “We incorporated everyone’s ideas.” “We all reminded each other to ‘not sweat the small stuff.’” “Frustrations boiled over to conversations with classmates that didn’t help our problems.” “The conflict was dealt with by people involved in talking it out.”

Idiosyncratic solutions Ideas related to fixing or

responding to problems

“Written rules to punish those who are late or lazy.” “We tried to correct agreeing too much by challenging things and playing “Devil’s advocate” better in the second case.” “Condescending tendencies—we gave people ‘time-outs’.” “We told that person if they made changes again without our permission that we’d kick them out.” “We instituted a time for “talking turns” which seemed to keep interruptions and arguments to a minimum.” “We told that person to stop rolling his eyes at us.” “We created a team calendar and made everyone put their classes and commitments on a week in advance. This made

scheduling a little easier.” “If you were late you have to buy everyone a Coke.”

Avoided or ignored Ideas about how groups

prevented and/or ignored conflict

“We kept conflicts underground!” “One member completely withdrew.” “We did whatever we could to avoid having to have team meetings.” “Some people just bit their tongue and went along with the majority in the interest of getting work done.” “One member was smart & made good contributions to the group, but was very insulting—so we listened to his

points and ignored the bad faces he made.” “Sometimes conflicts were not completely dealt with in essence of timing issue.” “Resolved conflict by setting up more well-defined processes to ensure group buy-in on timely basis.” “We arranged work so angry people could avoid each other.”

Rotating responsibilities Ideas about how to

allocate responsibility

“Who ever was the last “reviewer” of the paper had “veto” power about conflict.” “We assigned a member to be “on call” in case someone dropped the ball.” “Two people both wanted control. We gave it to the person who we thought had the best writing style and knew the

material the best.” “Each meeting had a team leader assigned and that person made the agenda.” “For example, on one case one guy could have done the case by himself—so we used him to double check the work

the rest of us did. So we learned and still did well.” “Because the members were not happy, we split the paper into 3 and divided the additional draft work.” “We spent more time in meetings rather than individually to work on cases and the result was very good.” “We gave everyone a chance to volunteer to do the work they wanted to do.”

178 BEHFAR, PETERSON, MANNIX, AND TROCHIM

provide additional evidence that our outcome patterns are meaningful. Examples of these norms are summarized in Table 4.

Consistently High or Increasing Performance and Satisfaction

This group of teams demonstrates the theoretical ideal of team- work in managing and preventing conflict.

Resolving relationship conflict. Only 1 of the teams in this category reported addressing relationship conflicts, which re- volved around negative non-verbal behavior or having to deal with an overly dominant personality. It reported using avoided/ignored to ensure bad behavior did not disrupt the group: “One member was smart & made good contributions to the group, but was very insulting—so we listened to his points and ignored the bad faces he made” (Team 26, Member 3).

Resolving process conflict. Seven teams, or 33% of the teams in this category, reported strategies to address process conflict, which primarily revolved around time management and the impact of domineering members on group process. There were two dom- inant management strategies: rotating responsibilities and discus- sion/compromise. Six of the 7 teams used both of these to either mitigate the negative impact of 1 member and/or to prevent any time management conflicts from escalating. For example, 1 team had a member with heavy interview commitments until mid- quarter— her team’s solution was “One member did most of the work later in the quarter, the rest of us pulled the load until then” (Team 34, Member 2). Another team was unsatisfied with the lateness and lack of commitment from 1 member and agreed to take turns ensuring work quality was up to team standards, “We had two people work on first assignment and made turns on next one” (Team 9, Member 2). These teams also used discussion and compromise to generate an understanding about how to mitigate conflicts before they escalated: “We met and voiced our concerns

regarding the quality and effort put into the paper (Team 38, Member 1)”; or “We talked about our schedules and then com- promised. One member gave up a weekend, another agreed to get up early one morning” (Team 34, Member 1).

Resolving task conflict. Fifteen teams, or 71% of the teams in this category, reported strategies for resolving task conflict. All 15 teams reported using three complementary strategies: compromise/ consensus, discuss/debate, and open communication. Teams used these strategies to secure solid understandings behind compromises/ group decisions by considering the pros and cons of different opin- ions, alternatives, and possible solutions. For example, “We put down what the two sides of the conflict were, then came back to them after we had nailed down everything we agreed on” (Team 6, Member 3), “We dealt with conflict by analyzing out each interpretation” (Team 31, Member 2), and “We talked thru the conflict by referencing facts that we collectively brainstormed” (Team 43, Member 3). These teams were able to reach compromise/consensus by considering evi- dence and convincing all members. For example, “We dealt with conflict by making sure everyone had a chance to fully explain their views and then went through a logical process of discussion to pick the most convincing argument” (Team 13, Member 1). They also identified that the nature of the conversation was not intertwined with negative emotion. For example, “The team members took the argu- ments positively and not personally” (Team 28, Member 1); and “No one was hard set on making their point; everyone was open” (Team 46, Member 1).

Consistently High/Increasing Performance, Consistently Low/Decreasing Satisfaction

These teams managed conflict with a focus on preventing con- tinuing or escalating conflict by putting more structure in the way individual team members worked together.

Table 3 Observed and Estimated Expected Frequencies for Changes in Norms by Outcome Pattern Category

Norms Trend

Team outcome pattern

Consistently high/increasing

performance and satisfaction

Consistently high/increasing

performance, consistently

low/decreasing satisfaction

Consistently low/decreasing

performance, consistently

high/increasing satisfaction

Consistently low/decreasing

performance and satisfaction Total

Time 1 Trouble Observed 0 6 0 7 13 Time 2 Trouble (Expected) (4.8) (2.5) (2.5) (3.2) (13)

Std. Residual �2.2 2.2 1.6 2.1 Time 1 Trouble Observed 4 0 1 0 5 Time 2 No Trouble (Expected) (1.8) (1.0) (1.0) (1.2) (5)

Std. Residual 1.6 1.0 0.0 1.1 Time 1 No Trouble Observed 5 5 5 4 19 Time 2 Trouble (Expected) (7.0) (3.7) (3.7) (4.7) (19)

Std. Residual 0.8 0.7 0.7 0.3 Time 1 No Trouble Observed 12 0 5 3 20 Time 2 No Trouble (Expected) (7.4) (3.9) (3.9) (4.9) (20)

Std. Residual 1.7 �2.0 0.6 0.9 Total Observed 21 11 11 14 57

(Expected) (21) (11) (11) (14) (57)

Note. Standardized residuals greater than or equal to 2 are shown in bold.

179CONFLICT MANAGEMENT IN TEAMS

Table 4 Examples of Team Norms

Team outcome pattern Team norms at Time 1 Team norms at Time 2

Consistently high/increasing performance and satisfaction

Team 7, Member 2

“Coming to meetings on time. Take nothing personally—constructive criticism only. Open communication of all ideas via email and at meetings. Make sure everyone on the same page—everyone understands why things are included and excluded in final paper. All functions of team members are interchangeable. Whoever wants to work on particular item just needs to ask—the group will be stronger this way.”

“Always think “we can do it.” No hesitation in changing opinion. Listen to other members carefully.”

Team 25, Member 1

“We try to have total participation and work to make sure that happens. We have made adjustments to remedy the problem of interrupting one another. We will put in at least a strong “B” and go for an “A.” We recognize that grading goals will evolve over the term. We set end times at the beginning of meetings and are pretty good at concluding at that mutually agreed upon time. We seem to balance one another pretty well. I’ve enjoyed working with this team. We have increased efficiency in the time we’ve worked together and I think we are pretty well aligned.”

“Respect, communication, and general approval.”

Consistently high/increasing performance, consistently low/decreasing satisfaction

Team 24, Member 3

“We have to collaborate, even if we don’t like it. We want an “A.” Our process (the process of getting the work done) is important and needs to be defined and agreed upon in advance.”

“We all tend to like to be in control. We have tried to balance control by moving responsibilities for writing from member to member for each draft. Our goals are aligned-we all want to do well. We communicate fairly well, but tend to take sides.”

Team 51, Member 3

“Despite the fact that 3 of us overruled the 4th person, the 4th person was unwilling to accept that. That person then went as far as adding information back into our paper that we had agreed to take out. This ended up compromising our grade. Very frustrating”

“3 of us get along well, work well together, and had a positive experience. The other member was always a challenge for us to deal with.”

Consistently low/decreasing performance, consistently high/increasing satisfaction

Team 52, Member 4

“A strong intent to seek consensus. Open and frank communication. Willingness to listen to as well as provide views on all group work.”

“Open ideas, trust and agreement.”

Team 65, Member 1

“Overall, the group worked well with one another. I believe we all contributed evenly to the preparation of the case and the write-up. One thing I would recommend is that all members of the group need to show up to meetings on time. Even being a few minutes late should not be tolerated.”

“We never spent too much time analyzing the material which I believed hurt us on some cases. It was always known that certain members of the group would be late.”

Consistently low/decreasing performance and satisfaction

Team 56, Member 1

“Two members were also easily frustrated and impatient though they had unrealistically short expectations of the amount of time required for a project.”

“We have a 2 vs. 2 situation in “How to do the project” viewpoints. 2 tend to cave, 1 tends to be overly aggressive, 1 tends to not be taken seriously.”

Team 59, Member 2

“There is friction about ideas that gets emotional, mostly caused by one person. We compromise to get past it.”

“We talk (and talk, and talk, and talk) until a consensus (or agreement that promises an end to talking) is reached.”

Note. All table entries are quotations from participants. Consistently low/decreasing performance, consistently high/increasing satisfaction had no significant pattern in the chi-square analysis.

180 BEHFAR, PETERSON, MANNIX, AND TROCHIM

Resolving relationship conflict. Six teams, or 55% of the teams in this category, reported addressing relationship conflicts. These conflicts revolved around recurrent clashes between domi- nant personalities and direct accusations that members made (e.g., stealing ideas, not caring about the team, and/or rude and conde- scending behavior). All 6 teams used strategies from the avoid/ ignore, idiosyncratic solutions, and discuss/debate categories. Taken together, these teams directly confronted relationship con- flict at the group level in order to prevent/avoid its reoccurrence. For example, teams managed escalating emotion by explicitly agreeing to overlook it: “We agreed to recognize when conflicts would help our ultimate goal–to get a better grade. It wasn’t easy, but helped everyone calm down” (Team 2, Member 2); and “Emo- tional outbursts were contained. We all knew the tension was there, we just agreed to not let it interfere” (Team 51, Member 2). While they agreed not to let emotion interfere, they also set up rules or created explicit expectations (idiosyncratic solutions) to help members adhere to this. For example, “When it got personal we ended the meeting. There was a cooling off period of about 30 minutes, and then we came back. This was harder when we procrastinated and didn’t have a spare 30 minutes” (Team 51, Member 1) and “We figured out how to avoid each other’s trigger points – for one it was meetings that drag on. For me it was being late AND unprepared” (Team 47, Member 2).

Resolving process conflict. Six teams, or 55% of the teams in this category, reported addressing process-related conflicts. These conflicts centered on disagreements about time manage- ment and workload distribution because of uneven contribu- tions or commitment from some members. Five of the 6 teams used two strategies together to address these conflicts: rotating responsibilities and idiosyncratic solutions. That is, they devel- oped systems to correct problems and/or rules to prevent prob- lems from occurring again. For example, “We figured out how to get the lame member to do something useful–take minutes, schedule meetings and talk to the TA” (Team 54, Member 2); “Conflict required empowering one person to step up to take charge” (Team 23, Member 1); and “We just agreed things needed to get better. Now you get punished ($10) or embar- rassed unless you perform” (Team 2, Member 2).

Resolving task conflict. Five teams, or 45% of the teams in this category, reported managing task-related conflicts. These conflicts centered on differences of opinions and ideas related to case analysis. Only 1 team managed emotionally charged task conflict with rules for “talking turns” (idiosyncratic solutions). The major- ity (4 of the 5 teams) used two strategies to manage their task conflicts: compromise and voting. Compromise, however, was associated with giving in or giving up one’s own opinion in the interest of getting work done. For example, “After considerable dialogue, the one individual agreed to go ahead with the project, but with an extreme amount of reservation” (Team 2, Member 1); and “Some people just said ‘OK, let’s just move on’” (Team 23, Member 2). The same 4 teams also reported using voting or majority rule to resolve task conflicts: “If you were on the losing side of a vote you just had to deal with it. Suck it up and move on. Be a team player” (Team 30, Member 3); and “We tended to let people express their points of view, but typically majority ruled at decision time” (Team 23, Member 3).

Consistently Low/Decreasing Performance, Consistently High/Increasing Satisfaction

These teams focused on relationships over task when managing conflicts.

Resolving relationship conflict. Only 2 out of 11 teams, or 18% of the teams in this category, reported experiencing relation- ship conflict. These 2 teams had 1 member that created emotional conflict in the teams. Both teams reported using strategies to avoid the conflict. For example, “We always divided work so those two people did not have to directly work together. The group avoided ever having to deal with that from beginning” (Team 35, Member 3) and “We decided to rally together–we all decided that that one person’s style was not a reflection of how the rest of us operate” (Team 45, Member 1). The remaining 9 teams in this category did not mention relationship conflict.

Resolving process conflict. Only 2 teams, or 18% of the teams in this category, reported managing process conflicts. These con- flicts centered on team meetings, such as members arriving late or when to schedule a meeting. Both teams resolved these conflicts with strategies from the rotating responsibilities category, which included changing/rotating work approach to foresee member con- flicts. For example, “We tried a new way of working together that better considered our busy schedules” (Team 65, Member 1); “We plotted out our busy times before the deadline and avoided meeting during those times” (Team 35, Member 2).

Resolving task conflict. Seven teams, or 64% of the teams in this category, reported managing task conflict. All of the teams used three strategies: open communication, discuss/debate, and compromise/consensus. Different from the teams described previ- ously, these teams did not have emotion intertwined in their task debates. Instead, there was not enough debate and too much focus on compromise: “We avoided conflict. I believe we were too quick to come to consensus” (Team 3, Member 2); and “Resolved by: We went through each solution to determine which was the easiest to defend with the information in the case” (Team 10, Member 3). Task conflicts were resolved by including all ideas rather than engaging in careful analysis.

Consistently Low/Decreasing Performance and Satisfaction

These teams had a hard time assessing the root cause of their performance and seemed willing to try anything to reduce the pain of their team experience.

Resolving relationship conflict. Five of the 14 teams, or 36% of the teams in this category, reported managing relationship conflict. These conflicts were severe: for example, “Tremendous personality clashes” (Team 53, Member 3) and “Members fought like crazy” (Team 59, Member 3). All 5 teams used open com- munication to address personality conflicts, and all 5 reported no success: “Unfortunately, personality differences created friction which never seemed to go away” (Team 44, Member 1); and “We did not resolve the conflict. We talked about expectations but no tangible improvements came out of that” (Team 53, Member 1).

Resolving process conflict. Eight teams, or 57% of the teams in this category, reported managing process conflicts. These con- flicts centered on expectations for time management and creating a process to work together. All of the teams used strategies from

181CONFLICT MANAGEMENT IN TEAMS

the rotating responsibilities and avoid/ignore categories, but the strategies reflect a lack of success: “We didn’t resolve much–still I don’t know at what point we went wrong. We did the best we could” (Team 66, Member 4); “We didn’t have the sense that everybody was satisfied with how the work was done (Team 41, Member 3); and “We all agreed to be more prepared and respon- sible, but it’s one thing to agree an another thing to actually DO IT” (Team 67, Member 5). The strategies from the rotating re- sponsibilities all described “rotation” of approach to work: “We tried brainstorming, we tried devil’s advocate, we tried it all” (Team 66, Member 1); and “We tried different ways/approaches for each case” (Team 42, Member 1).

Resolving task conflict. Nine teams, or 64% of the teams in this category, reported managing task conflict. Eight of these teams reported using the same two strategies: discuss/debate and compromise/consensus. However, the compromise/consensus units reflected a pattern of “giving in” to avoid unpleasant expe- riences: “Mostly I just said “whatever” which probably wasn’t great team behavior” (Team 66, Member 3); and “One member was domineering and a control freak. Most of the time the group gave in to the dominant member” (Team 59, Member 1). The discuss/debate units reflected a focus on issues tangential to the actual work: “Conflict continued about what constituted an anal- ysis versus what was simply a retelling of the case” (Team 56, Member 1) and “Although majority of the time was spent on determining how to convince one or more members to the ideas (Team 19, Member 4).

Summary

These results and differences between the four outcome trends are summarized in Table 5. Overall, the results of the summary analysis suggest that while teams may have reported using strat- egies from the same conflict resolution categories, the way that they applied those strategies to resolve different conflicts was often associated with very different patterns of change in perfor- mance and satisfaction.

Discussion

The teams in this study all started from the same point—the groups were newly formed, endowed largely with the same re- sources, had the same task assignments and performance evalua- tion criteria, and were working under the same time constraints. Yet, even under these very similar operating conditions, they developed and applied conflict resolution strategies in very differ- ent ways with different results. Our results suggest a number of implications for the literature on conflict in teams.

Theoretical Contributions

There are two interrelated theoretical contributions this study makes. The first is to provide detail about how teams manage task, relationship, and process conflicts as well as the performance and satisfaction tradeoffs associated with choices in conflict resolution strategies. For example, our results suggest that majority rule voting is associated with consistently high and increasing task performance but consistently low and decreasing member satisfac- tion. We reveal these results by looking at these issues through a

surprisingly underutilized frame of reference. That is, we look at conflict resolution strategies at the group rather than individual level (i.e., individual conflict styles), and we induce multiple team-level strategies rather than look at or impose a single strategy (e.g., a cooperative frame). In that sense, we take a closer or more fine-grained look at team-level conflict resolution strategies than has been done in the past. Interestingly, our participant-driven categorization of team conflict management does loosely map onto current individual-level conflict management typologies (e.g., Blake & Mouton, 1964; Pruitt & Rubin, 1986; Rahim, 1983; Ruble & Thomas, 1976), and the results of the summary analysis are consistent with the underlying theory. Our results, for example, are consistent with existing conclusions about the benefits of integra- tive and collaborative approaches to managing conflict and the drawbacks of contending and avoiding approaches (De Dreu & VanVianen, 2001; Pruitt & Rubin, 1986; Tjosvold et al., 2003). The interesting addition to current literature from these results comes from comparing management of the three types of conflict between different outcome patterns. We propose, therefore, that, rather than by directly mapping existing individual-level conflict management “styles” onto our findings, the results of this study can be better characterized by Hackman and Morris’s (1975) three criteria for team viability: (a) The team must meet the expectations of those who receive their work (i.e., performance); (b) the team needs to satisfy the individual needs of members in the group experience (i.e., individual member satisfaction), and (c) the pro- cess the group uses (e.g., conflict management tactics) must en- hance its ability to work together in the future. We do, however, note where there is overlap between the individual conflict man- agement styles literature and our own findings. We also discuss the results of the summary analysis in terms of how differences in conflict management focus reflect (or not) shared governance of team process (criterion #3) and how these are associated with outcome patterns (criteria #1 and #2). In short, we identify overlap with existing individual conflict management style typologies but do not try to force the categories where there is not complete conceptual agreement across levels of analysis.

The teams that were consistently high/increasing in performance and satisfaction represent fulfillment of all three team viability criteria. The focus these teams developed in their conflict man- agement strategies was on “equity,” or finding an appropriate (not necessarily equal) way for all members to contribute given their constraints (a mix of accommodating, collaborating, and compro- mising). They were most likely able to achieve performance goals and high individual member satisfaction by proactively securing solid understanding of the compromises and trade-offs both the team and individual members made in planning how to prevent potentially destructive conflicts. Stated in the language of the commonly cited “dual concerns” approach—these teams had both a concern for their task and a concern for integrating the interests of individual members. They took this approach to managing both task and process conflicts. Their norms at Time 1 and Time 2 were coded as functional, and they seemed to establish clear expecta- tions and shared criteria for selection (e.g., of ideas and work distribution) if there was disagreement.

In contrast, the teams that were consistently high/increasing in performance but low/decreasing in satisfaction had more of a “rules” focus toward conflict management (a mix of competing/ forcing and compromising). They were more focused on reacting

182 BEHFAR, PETERSON, MANNIX, AND TROCHIM

to conflicts and structuring or creating rules to avoid the continued disruption in meeting performance goals than on preemptively planning to prevent such disruptions. This approach created a focus on task over integrating individual interests and is reflected in their majority-rules approach to task conflict, in the boundary conditions they set for managing relationship conflict, and in the arbitration approach they took for process conflict. As such, the disruptive impact of conflict was “managed” but, as reflected in

the team norms accounts coded as troubled, the residual tension among members never completely subsided (they in fact agreed to “overlook” the interpersonal tension).

Reciprocally to these teams, the low/decreasing performance and high/increasing satisfaction teams developed an “equality” focus (very similar to an accommodating style) where complete accommodation of individuals took the place of rules or criteria for selection. While conclusions can be drawn from these data only

Table 5 Summary of Conflict Resolution Strategies Organized by Outcome Category

Type of conflict

Team patterns in applying conflict management strategies

Consistently high/increasing perfor- mance and satisfaction

Consistently high/increasing performance, consistently

low/decreasing satisfaction

Consistently low/decreasing perfor-

mance, consistently high/increasing

satisfaction Consistently low/decreasing performance and satisfaction

(21 teams, 55 conflict management units)

(11 teams, 33 conflict management units)

(11 teams, 46 conflict management units)

(14 teams, 40 conflict management units)

Task Used discuss/debate and open communication to reach compromise/ consensus.

Used compromise/consensus and voting.

Used discuss/debate and open communication to reach compromise/ consensus.

Used discuss/debate and compromise/consensus.

Non-emotional, fact-driven discussions helped team members understand how the group reached consensus. Conflicting views were explored and consensus was reached based on evidence that convinced all members.

Compromise was associated with a “majority rules” focus—if one member did not agree he/she conceded to the consensus of the majority. When this concession did not occur, these teams used more formal voting strategies to demonstrate consensus.

Discussion of different ideas was amicable and consensus centered on finding a way to incorporate all member’s ideas rather than engaging in debate to select the best ideas.

Consensus was associated with either the group “giving in” to a dominant member or by individuals “giving in” because they were tired of arguing.

Relationship Inconclusive: Only one team reported managing relationship conflict.

Used discuss/debate and idiosyncratic solutions to avoid/ignore conflicts.

These teams explicitly discussed the impact of relationship conflict on the team and agreed it should not be allowed to impede performance. Members agreed to rules or procedures that enforced this agreement (e.g., cooling off periods).

Inconclusive: Only two teams reported managing relationship conflict.

Used open communication. These teams experienced

severe interpersonal conflict, which members openly discussed. However, all teams reported this did not change or eliminate these conflicts.

Process Used discussion/debate with compromise/consensus to employ rotating responsibilities.

These teams forecasted scheduling and commitment conflicts and discussed quality of work concerns early. Work assignments were made to prevent predicted problems from occurring. Members made compromises in their schedules and work assignments to ensure quality of work.

Used rotating responsibilities and idiosyncratic solutions.

They created rules or procedures to address previous time management

or member contribution issues. Members agreed how to share (rotate) responsibility for correcting problems (e.g., empowering one member as an arbitrator, monetary penalties).

Inconclusive: Only two teams reported managing process conflict.

Used rotating, responsibilities and avoided/ignored.

Rotating Responsibilities centered on “rotation” of work approaches or task strategies intended to correct previous process and perfor- mance problems. Units coded in the avoided/ignored category all reflected a lack of resolution.

183CONFLICT MANAGEMENT IN TEAMS

about task conflict management strategies, their strategies were heavily classified by participants as “avoidant,” as they avoided difficult conversations in the interest of preserving interpersonal relationships and never really seemed to face the difficult task of integrating various interests and perspectives to create a superior team product. Thus, they maintained interpersonal harmony at the cost of relatively poor task performance.

Finally, the teams that were low/decreasing in performance and satisfaction took an ad hoc approach to managing conflict—roles were never clear; the root cause of problems was not identified nor successfully corrected; there was no central/coherent conflict man- agement strategy; and communication came at a high cost. In the traditional conflict management typologies, these teams represent a double-edged mix of competing/forcing and avoiding styles. While these teams were willing to directly confront and argue about their interests, this process (as reflected in the team norms accounts) became an obstacle to working together effectively. As a result, teams tended to resign or make superficial accommoda- tions to get past the conflict, which neither improved their task focus nor integrated the interests of members. For example, they resolved task conflict by either giving into a dominant member or spending too much time trying to convince 1 member; their dis- cussion of relationship conflict usually escalated rather than re- solved the problem; and their attempts to resolve process conflict

centered on more superficial fixes such as trial and error with different processes. This is consistent with previous findings about teams that have difficulty establishing stable and functional con- flict resolution approaches (see Argyris, 1982; Jehn & Mannix, 2001).

The second theoretical contribution of our work is to develop group-level theory about conflict resolution strategies. Taking a step back from the specific conflict resolution tactics we found, an underlying pattern emerges— one of two broad dimensions that describe how the matching of conflict resolution strategies with conflict create a pattern of group outcomes. We show this in Figure 2. Notice, for example, that the teams with mixed outcome changes during the performance episode (22 teams in this sample) demon- strate how a different conflict management focus, either creating rules in reaction to conflict versus accommodating all members to prevent it, can transform the impact of conflict. In the rules- focused teams, for example, the negative impact of process and relationship conflict was suppressed by their conflict management approach and enabled the group to engage in “healthy” debate- driven task conflict. However, a lack of focus on integrating individual interests into the team solution may have caused satis- faction to suffer, perhaps because peers had to enforce rules, and without rules members did not feel they could count on each other to behave in the best interests of the team. In the equality-focused

Preemptive Strategies

Conflict resolution strategies planned to preempt negative effects of conflict

Reactive Strategies

Conflict resolution strategies applied in reaction to existing problems

Pluralistic Strategies

Conflict resolution strategies that apply to all

in the group

Increasing and Consistently High Performance/ Increasing and Consistently High Performance/

Increasing and Consistently High Satisfaction Decreasing and Consistently Low Satisfaction

Resolution Focus: Equity Resolution Focus: Creating Explicit Rules

Foreseeing problems and preemptively Reacting to previous disruptions by restructuring and

clarifying expectations, including: organizing to eliminate disruptions, including:

Work assignments based on skill Written rules & punishments

Forecasting scheduling & work-load problems Majority rules at decision time

Understanding the reasons behind compromises Arbitration approach to conflict

Focusing on content over delivery style

11 teams, 33 statements 21 teams, 55 statements

Decreasing and Consistently Low Performance/

Decreasing and Consistently Low Satisfaction

Resolution Focus: Adhocracy

Reacting to previous problems by focusing on minimizing individual misery, including:

Divide and conquer Avoiding debate, choose easy solution

Trial and error to correct process Avoiding group meetings

Decreasing and Consistently Low Performance/

Increasing and Consistently High Satisfaction

Resolution Focus: Equality

Anticipating how group decisions impact individual feelings, including:

Work assignments based on volunteersParticularistic Strategies

Conflict resolution strategies aimed at satisfying or

containing individual members

In place of analysis, include all ideas Strong focus on cohesion

Figure 2. The role of conflict resolution in predicting group outcomes.

184 BEHFAR, PETERSON, MANNIX, AND TROCHIM

teams, there was very little mention of relationship or process conflict as their desire to suppress disagreement influenced the decisions they made in structuring their process. So, while pre- emptive prevention of conflict made team members feel good, their approach did not address bigger issues underlying task achievement (e.g., extracting member expertise).

Overall, we would argue that these results suggest a theoretical melding of findings from the conflict management literature and Hackman and Morris’s (1975) widely cited team viability model. Figure 2, the summary of our key findings, could also be seen as an updated version of this model. These results can be summarized by two striking and interrelated differences that developed be- tween teams in this study. Along the vertical axis, we categorize conflict management as particularistic versus pluralistic. Groups are considered pluralistic if they take a “whole-group” perspective in establishing processes that apply to everyone and are for the benefit of the task performance. A particularistic perspective, on the other hand, is one in which decisions are made to contain or respond to a particular person or situational conflict rather than to enhance the big picture of task performance. Teams with consis- tently high or increasing performance were more likely to have utilized group-level or pluralistic conflict management strategies (e.g., rules that clarify expectations and apply to all group mem- bers), whereas teams with consistently low/decreasing perfor- mance were more particularistic and focused on solutions to ad- dress individual negative reactions to conflict (e.g., dividing work according to “turns” rather than expertise so as not to upset individuals). Pluralistic conflict management strategies tended to be better at correcting or preventing the root cause of the conflict from continuing to impact the group.

Along the horizontal axis, we categorize conflict management as reactive versus preemptive (cf. Marks et al., 2001; Weingart, 1997). Groups are considered preemptive if they make decisions about group resources (time, member skills, materials, etc.) in a way that integrates individual interests by identifying issues and creating solutions to potential conflicts before they arise. A reac- tive orientation, on the other hand, is characterized by backward- looking decision making—with previous process mistakes and/or conflict experiences having higher salience in making decisions about how the group should work together in the future than have individual interests. This framework suggests that these strategies may be indicators of conflict management strategies that may or may not break negative spirals before they begin (Brett, Shapiro, & Lytle, 1998).

Overall, teams in the consistently high/increasing outcomes are the ones on a clearly positive trajectory and seem most likely to succeed over time. Figure 2 demonstrates how the intersection of viability criterion number one (performance) and two (satisfaction) affect a group’s orientation toward the third (operationalized as conflict management in this study). That is, in order for a team to continue to adapt its processes to achieve performance expecta- tions, it must also manage the social consequences (primarily the relationship and process conflicts) of group behaviors that de- crease individual motivation to constructively contribute. Thus, our argument suggests why Figure 2 may represent an updated group viability model summarizing how different applications of conflict resolution strategies can be used to understand why group processes may or may not enhance prospective ability to work together at any given point in time. In sum, we have argued that the

way a team manages conflict resolution and adapts its process is of paramount importance to understanding why teams struggle to sustain high performance and/or high individual member satisfac- tion over the longer term (Hackman & Morris, 1975; Homans, 1950; McGrath, 1964; see also Cordery et al., 1991; Goodman et al., 1988; Guzzo, 1982; Howell et al., 1990; Marks et al., 2001; Moorhead et al., 1998; Steiner, 1972). Our results suggest that it is not just the type of conflict a team experiences that matters; rather, how well conflict resolution strategies address a team-level bal- ance between task and affect management is what yields team viability. An imbalance of one kind or another is associated with a threat to viability.

Limitations and Future Directions

Like all single studies, there are reasons for exercising some caution in generalizing our results. Our study has a number of interrelated strengths and limitations that are related to the sample chosen (i.e., an MBA classroom). First, the categories of resolution strategies used by MBA students may not apply to all work settings. We believe, however, that the external pressures, conse- quences, and intensity of their work together can provide a rea- sonable estimation for an organizational outcome to be investi- gated in future research. This study is only a first step in sketching an overall picture of relationships among conflict resolution strat- egies, conflict behaviors, and group outcomes. It was conducted to better understand qualitative differences in conflict management strategies in newly forming groups embedded in a relatively sim- ple organizational system. The reward structure and incentives in this sample were relatively standardized compared with an orga- nization with political and economic pressures.

The second limitation of our study revolves around group de- sign. Our groups had complete autonomy to self-manage, whereas most organizational groups have team leaders or mangers that are likely to provide some direction or structure. The teams literature has paid scant attention to the influence of team design, authority, and feedback on group process (for an exception, see Hackman & Wageman, 2005). Although leadership is also often recognized as a structural variable (e.g., Gladstein, 1984), like most group studies that use student groups (autonomous by design), this study does not account or control for the influence of a manager on group process. Intervention of a legitimate authority and performance feedback can have a significant impact on how group process evolves.

The third limitation of our study revolves around the fact that performance feedback was given in the form of a grade rather than a discussion of performance as might be done in an organization. We cannot be entirely certain if teams were revising their process management strategies in reaction to previous process experiences or their grades since it was not observed nor discussed with them. So, although our results specifically suggest that performance feedback and benchmarking task accomplishment are important forces in shaping group process, our results should be interpreted with some caution because they were not delivered in a typical organizational format. The information given to a group through periodic feedback discussion and the legitimate authority of an appointed leader can be important mechanisms for helping a group self-correct a process routine (cf. Argyris, 1985). In addition, the ability to “blame” an external manager for critical feedback or an

185CONFLICT MANAGEMENT IN TEAMS

outcome, rather than hearing it from and sharing responsibility with peers, may have a different impact on member satisfaction.

Finally, this study investigated only one performance episode. Future investigations should investigate multiple waves of work– feedback cycles to better assess causality between strategies and outcomes. The results of this study can indicate only the initial trajectory a team will have over time. Future research should also investigate how teams might change their process focus over time as well as how different types of task assignments (e.g., creative vs. problem-solving) may require a different conflict resolution focus, and it should more directly document the impact of accu- mulating performance and process feedback on evolving conflict management approaches. While functional background, gender, and so on, were randomly assigned in this sample, it is not likely to be so in the real organizations. Care needs to be taken to better control for covariates that might magnify between-teams differ- ences (e.g., Cropanzano, Aguinis, Schminke, & Denham, 1999).

All of these limitations should not be overstated, however. Our study investigates autonomous groups that all started from exactly the same place and with the same resources in terms of time and talent. Like all organizational teams, they were under strong work pressure to produce outputs for others who evaluate their work where those evaluations of the work are consequential. Even so, we saw striking differences in the conflict-management strategies these teams employed that were associated with changes in satis- faction and performance over time. We have focused here, how- ever, only on conflict and conflict resolution strategies. Other scholars have demonstrated related links between conflict and additional group-level variables such as diversity, efficacy, cohe- sions, and emotion (e.g., Ancona & Caldwell, 1992; De Dreu & VanVianen, 2001; Jehn, 1995, 1997; Jehn, Northcraft, & Neale, 1999; Jehn & Shah, 1997). More research needs to be done to fully understand the complex set of relationships between all of these variables. Most importantly, however, this study suggests that conflict resolution strategies in groups ought to receive greater attention in predicting team performance than has been the case to date.

References

Agresti, A. (1990). Categorical data analysis. New York: Wiley. Amason, A. C., & Sapienza, H. (1997). The effects of top management

team size and interaction norms on cognitive and affective conflict. Journal of Management, 23, 495–516.

Ancona, D., & Caldwell, D. (1992). Bridging the boundary: External activity and performance in organizational teams. Administrative Sci- ence Quarterly, 37, 634 – 665.

Ancona, D., & Chong, C. (1996). Entrainment: Pace, cycle, and rhythm in organizational behavior. Research in Organizational Behavior, 18, 251– 284.

Argyris, C. (1982). Reasoning, learning and action: Individual and orga- nizational. San Francisco: Jossey-Bass.

Argyris, C. (1985). Strategy, change, and defensive routines. Marshfield, MA: Pitman.

Arrow, H., McGrath, J., & Berdahl, J. (2000). Small groups as complex systems: Formation, coordination, development, and adaptation. Thou- sand Oaks, CA: Sage.

Arrow, H., Poole, M. S., Henry, K. B., Wheelan, S., & Moreland, R. (2004). Time, change, and development: The temporal perspective on groups. Small Group Research, 35, 73–105.

Beekun, R. (1989). Assessing the effectiveness of sociotechnical interven- tions: Antidote or fad? Human Relations, 42, 877– 897.

Bettenhausen, K., & Murnighan, J. (1985). The emergence of norms in competitive decision making groups. Administrative Science Quarterly, 30, 350 –372.

Blake, R., & Mouton, J. (1964). The managerial grid. Houston, TX: Gulf. Brett, J. (2001). Negotiating globally: How to negotiate deals, resolve

disputes, and make decisions across cultural boundaries. San Francisco: Jossey-Bass.

Brett, J., Shapiro, D., & Lytle, A. (1998). Breaking the bonds of reciprocity in negotiation. Academy of Management Journal, 41, 410 – 424.

Cohen, S., & Ledford, G. (1994). The effectiveness of self-managing teams: A quasi-experiment. Human Relations, 47, 13– 43.

Cordery, J., Mueller, W., & Smith, L. M. (1991). Attitudinal and behav- ioral effects of autonomous group working: A longitudinal field study. Academy of Management Journal, 34, 464 – 476.

Cropanzano, R., Aguinis, H., Schminke, M., & Denham, D. (1999). Dis- putant reactions to managerial conflict resolution tactics: A comparison among Argentina, the Dominican Republic, Mexico, and the United States. Group and Organization Management, 24(2), 124 –154.

Cummings, T., & Griggs, W. (1977). Worker reactions to autonomous work groups— conditions for functioning, differential effects, and indi- vidual differences. Organization and Administrative Sciences, 7, 87– 100.

De Dreu, C. (1997). Productive conflict: The importance of conflict man- agement and conflict issues. In C. de Drew & E. Van de Vliert (Eds.), Using conflict in organizations (pp. 9 –22). London: Sage.

De Dreu, C. (2006). When too little or too much hurts: Evidence for a curvilinear relationship between task conflict and innovation in teams. Journal of Management, 32(1), 83–107.

De Dreu, C., & VanVianen, A. (2001). Managing relationship conflict and the effectiveness of organizational teams. Journal of Organizational Behavior, 22, 309 –328.

De Dreu, C., & Weingart, L. (2003). Task versus relationship conflict, team performance, and team member satisfaction. Journal of Applied Psy- chology, 88, 741–749.

Desivilya, H., & Yagil, D. (2005). The role of emotions in conflict management: The case of work teams. International Journal of Conflict Management, 16, 55– 69.

Edelmann, R. (1993). Interpersonal conflicts at work. Exeter, United Kingdom: British Psychological Society.

Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44, 350 –383.

Feldman, D. (1984). The development and enforcement of group norms. Academy of Management Review, 9, 47–53.

Fine, G., & Elsbach, K. (2000). Ethnography and experiment in social psychological theory building: Tactics for integrating qualitative field data with quantitative lab data. Journal of Experimental Social Psychol- ogy, 36, 51–76.

Fleiss, J. (1981). The measurement of interrater agreement. In Statistical methods for rates and proportions (2nd ed., pp. 212–236). New York: Wiley.

Folger, J., Poole, M., & Stutman, R. (2001). Working through conflict: Strategies for relationships, groups, and organizations (5th ed.). New York: Longman.

Geer, J. G. (1991). Do open-ended questions measure ‘salient’ issues? Public Opinion Quarterly, 55, 360 –370.

Gladstein, D. (1984). Groups in context: A model of task group effective- ness. Administrative Science Quarterly, 29, 499 –517.

Goodman, P., Devadas, R., & Hughson, T. G. (1988). Groups and produc- tivity: Analyzing the effectiveness of self-managing teams. In J. Camp- bell, R. Campbell, & Associates (Eds.), Productivity in organizations: New perspectives from industrial and organizational psychology (pp. 295–327). San Francisco: Jossey-Bass.

186 BEHFAR, PETERSON, MANNIX, AND TROCHIM

Guzzo, R. (Ed.). (1982). Improving group decision making in organiza- tions: Approaches from theory and research. New York: Academic Press.

Guzzo, R., & Shea, G. (1992). Group performance and intergroup relations in organizations. In M. Dunnette & L. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., Vol. 3, pp. 269 –313). Palo Alto, CA: Consulting Psychologists Press.

Hackman, J. R. (Ed.). (1990). Groups that work (and those that don’t): Creating conditions for effective teamwork. San Francisco: Jossey-Bass.

Hackman, J., & Morris, C. (1975). Group tasks, group interaction process, and group performance effectiveness: A review and proposed integra- tion. Advances in Experimental Social Psychology, 8, 45–99.

Hackman, J. R., & Wageman, R. (2005). A theory of team coaching. Academy of Management Review, 30, 269 –287.

Hinkin, T., & Tracey, J. B. (1999). An analysis of variance approach to content validation. Organizational Research Methods, 2, 175–186.

Homans, G. (1950). The human group. New York: Harcourt, Brace and Company.

Howell, J., Bowen, D., Dorfman, P., & Podskaoff, P. (1990). Substitutes for leadership: Effective alternatives to ineffective leadership. Organi- zational Dynamics, 19, 21–38.

Ilgen, D. J. H., John, M., & Jundt, D. (2005). Teams in organizations: From input–process– output models to IMOI models. Annual Review of Psy- chology, 56, 517–543.

Jackson, K., & Trochim, W. (2002). Concept mapping as an alternative approach for the analysis of open-ended survey questions. Organiza- tional Research Methods, 5, 307–336.

James, L. (1982). Aggregation bias in estimates of perceptual agreement. Journal of Applied Psychology, 67, 219 –229.

James, L., Demaree, R., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85–98.

Jehn, K. (1995). A multimethod examination of the benefits and detriments of intragroup conflict. Administrative Science Quarterly, 40, 256 –282.

Jehn, K. (1997). A qualitative analysis of conflict types and dimensions in organizational groups. Administrative Science Quarterly, 42, 530 –557.

Jehn, K., & Bendersky, C. (2003). Intragroup conflict in organizations: A contingency perspective on the conflict– outcome relationship. Research in Organizational Behavior, 24, 187–242.

Jehn, K., & Mannix, E. (2001). The dynamic nature of conflict: A longi- tudinal study of intragroup conflict and group performance. Academy of Management Journal, 44, 238 –251.

Jehn, K., Northcraft, G., & Neale, M. (1999). Why differences make a difference: A field study of diversity, conflict, and performance in workgroups. Administrative Science Quarterly, 44, 741–763.

Jehn, K., & Shah, P. (1997). Interpersonal relationships and task perfor- mance: An examination of mediating processes in friendship and ac- quaintance groups. Journal of Personality and Social Psychology, 72, 775–790.

Jick, T. (1979). Mixing qualitative and quantitative methods: Triangulation in action. Administrative Science Quarterly, 24, 602– 611.

Klein, K., Bliese, P., Kozlowski, S., Dansereau, F., Gavin, M., Griffin, M., et al. (2000). Multilevel analytical techniques: Commonalities, differ- ences, and continuing questions. In K. Klein & S. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 512–553). San Francisco: Jossey- Bass.

Kraut, A. (Ed.). (1996). Organizational surveys: Tools for assessment and change. San Francisco: Jossey-Bass.

Kruskal, J., & Wish, M. (1978). Multidimensional scaling. Beverly Hills, CA: Sage.

Kuhn, T., & Poole, M. S. (2000). Do conflict management styles affect group decision making? Evidence from a longitudinal field study. Hu- man Communication Research, 26, 558 –590.

Lance, C., Butts, M., & Michels, L. (2006). The sources of four commonly reported cutoff criteria. Organizational Research Methods, 9(2), 202– 220.

Lind, E. A., & Earley, P. C. (1992). Procedural justice and culture. International Journal of Psychology, 27, 227–242.

Lind, E. A., & Tyler, T. (1988). The social psychology of procedural justice. New York: Plenum.

Lindsley, D., Brass, D., & Thomas, J. (1995). Efficacy–performance spi- rals: A multilevel perspective. Academy of Management Review, 20, 645– 678.

Loyd, D., Thompson, L., & Kern, M. (2005). Classroom research: Bridging the ivory divide. Academy of Management Learning and Education, 4, 8 –21.

Marks, M., Mathieu, J., & Zaccaro, S. (2001). A temporally based frame- work and taxonomy of team process. Academy of Management Review, 26, 356 –376.

Mathieu, J., & Schulze, W. (2006). The influence of team knowledge and formal plans on episodic team process–performance relationships. Acad- emy of Management Journal, 49, 605– 619.

McGrath, J. (1964). Social psychology: A brief introduction. New York: Holt, Rinehart, and Winston.

McGrath, J. (1991). Time, interaction, and performance (TIP): A theory of groups. Small Group Research, 22, 147–174.

Miles, M., & Huberman, M. (1994). Qualitative data analysis: An ex- panded sourcebook (2nd ed.). Thousand Oaks, CA: Sage.

Moorhead, G., Neck, C., & West, M. (1998). The tendency toward defec- tive decision making within self-managing teams: The relevance of groupthink for the 21st century. Organizational Behavior and Human Decision Processes, 73, 327–351.

Morrill, C., & Thomas, K. (1992). Organizational conflict-management as disputing process: The problem of social escalation. Human Communi- cation Research, 18(3), 400 – 428.

Nunnally, J., & Bernstein, I. (1994). Psychometric theory (3rd ed.). New York: McGraw Hill.

Okhuysen, G., & Eisenhardt, K. (2002). Integrating knowledge in groups: How formal interventions enable flexibility. Organization Science, 13, 370 –386.

Pasmore, W., Francis, C., & Haldeman, J. (1982). Sociotechnical systems: A North American reflection on empirical studies of the seventies. Human Relations, 35, 1179 –1204.

Peterson, R. S. (1997). A directive leadership style in group decision making is both virtue and vice: Evidence from elite and experimental groups. Journal of Personality and Social Psychology, 72, 1107–1121.

Peterson, R., & Behfar, K. J. (2003). The dynamic relationship between performance feedback, trust, and conflict in groups: A longitudinal study. Organizational Behavior and Human Decision Processes, 92, 102–112.

Pondy, L. (1992). Reflections on organizational conflict. Journal of Orga- nizational Behavior, 13, 257–261.

Poole, M. S., Siebold, D., & McPhee, R. (1996). The structuration of group decisions. In R. Hirokawa & M. S. Poole (Eds.), Communication and group decision making (pp. 114 –146). Thousand Oaks, CA: Sage.

Pruitt, D., & Rubin, J. (1986). Social conflict: Escalation, stalemate, and settlement. New York: Random House.

Putnam, L., & Poole, M. S. (1987). Conflict and negotiation. In F. Jablin, L. Putnam, K. Roberts, & L. Porter (Eds.), Handbook of organizational communication: An interdisciplinary perspective (pp. 549 –599). New- bury Park, CA: Sage.

Rahim, M. A. (1983). A measure of styles of handling interpersonal conflict. Academy of Management Journal, 26, 368 –376.

Rahim, M. A. (2002). Toward a theory of managing organizational con- flict. International Journal of Conflict Management, 13, 206 –235.

Ruble, T., & Thomas, K. (1976). Support for a two-dimensional model of

187CONFLICT MANAGEMENT IN TEAMS

conflict behavior. Organizational Behavior and Human Performance, 16, 143–155.

Simons, T. L., & Peterson, R. S. (2000). Task conflict and relationship conflict in top management teams: The pivotal role of intragroup trust. Journal of Applied Psychology, 85, 102–111.

Steiner, I. (1972). Group process and productivity. New York: Academic Press.

Thomas, K. (1992). Conflict and negotiation processes in organizations. In M. Dunnette & L. Hough (Eds.), Handbook of industrial and organiza- tional psychology (Vol. 3, pp. 651–717). Palo Alto, CA: Consulting Psychologists Press.

Tinsley, C. (2001). How negotiators get to yes: Predicting the constellation of strategies used across cultures to negotiate conflict. Journal of Ap- plied Psychology, 86, 583–593.

Tjosvold, D. (1991). The conflict-positive organization. Boston: Addison- Wesley

Tjosvold, D., Hui, C., & Yu, Z. (2003). Conflict management and task reflexivity for team in-role and extra-role performance in China. Inter- national Journal of Conflict Management, 14, 141–163.

Vancouver, J. B. (2000). Self-regulation in organizational settings: A tale of two paradigms. In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.),

Handbook of self-regulation (pp. 303–336). San Diego, CA: Academic Press.

Wall, J., & Callister, R. (1995). Conflict and its management. Journal of Management, 21, 515–558.

Wall, T., Kemp, N., Jackson, P., & Clegg, C. (1986). Outcomes of autonomous workgroups: A long-term field experiment. Academy of Management Journal, 29, 280 –304.

Weingart, L. (1992). Impact of group goals, task component complexity, effort, and planning on group performance. Journal of Applied Psychol- ogy, 77, 682– 693.

Weingart, L. (1997). How did they do that? The ways and means of studying group process. Research in Organizational Behavior, 19, 189 – 239.

Williams, K., & O’Reilly, C. (1988). Demography and diversity in orga- nizations: A review of 40 years of research. Research in Organizational Behavior, 20, 77–140.

Received January 19, 2007 Revision received June 15, 2007

Accepted July 9, 2007 �

Members of Underrepresented Groups: Reviewers for Journal Manuscripts Wanted

If you are interested in reviewing manuscripts for APA journals, the APA Publications and Communications Board would like to invite your participation. Manuscript reviewers are vital to the publications process. As a reviewer, you would gain valuable experience in publishing. The P&C Board is particularly interested in encouraging members of underrepresented groups to participate more in this process.

If you are interested in reviewing manuscripts, please write to the address below. Please note the following important points:

• To be selected as a reviewer, you must have published articles in peer-reviewed journals. The experience of publishing provides a reviewer with the basis for preparing a thorough, objective review.

• To be selected, it is critical to be a regular reader of the five to six empirical journals that are most central to the area or journal for which you would like to review. Current knowledge of recently published research provides a reviewer with the knowledge base to evaluate a new submission within the context of existing research.

• To select the appropriate reviewers for each manuscript, the editor needs detailed information. Please include with your letter your vita. In the letter, please identify which APA journal(s) you are interested in, and describe your area of expertise. Be as specific as possible. For example, “social psychology” is not sufficient—you would need to specify “social cognition” or “attitude change” as well.

• Reviewing a manuscript takes time (1– 4 hours per manuscript reviewed). If you are selected to review a manuscript, be prepared to invest the necessary time to evaluate the manuscript thoroughly.

Write to Journals Office, American Psychological Association, 750 First Street, NE, Washington, DC 20002-4242.

188 BEHFAR, PETERSON, MANNIX, AND TROCHIM

Correction to Behfar et al. (2008)

In the article “The Critical Role of Conflict Resolution in Teams: A Close Look at the Links Between Conflict Type, Conflict Management Strategies, and Team Outcomes,” by Kristin J. Behfar, Randall S. Peterson, Elizabeth A. Mannix, and William M. K. Trochim (Journal of Applied Psychology, 2008, Vol. 93, No. 1, pp. 170 –188), Figure 2 on p. 184 (Theoretical Contributions section) is missing information about the numbers of teams and statements for the two groups using particularlistic strategies. For the decreasing and consistently low performance/decreasing and consistently low satisfaction group, there were 14 teams and 40 statements; for the decreasing and consistently low performance/increasing and consistently high satisfaction group, there were 11 teams and 46 statements.