A Helping Hand? The Moderating Role of Leaders' Conflict Management Behavior on the Conflict–Stress Relationship of Employees

Conflict, which is defined as a process between two individuals that arises when one party feels obstructed or irritated by the other (Van de Vliert 1997), occurs frequently among employees (e.g., Wall and Callister 1995). Individuals experiencing conflict often feel anxiety, frustration, and tension (Spector, Chen, and O'Connell 2000; Spector and Bruk‐Lee 2008), and conflict has been found to negatively affect employees' job satisfaction and performance (De Dreu and Weingart 2003). Obviously, organizations need to manage employee conflict effectively to minimize these negative consequences. The importance of conflict management for well‐being, including employee satisfaction (Behfar et al. 2008) or employee stress (Friedman et al. 2000), supports this notion.

Generally, a leader's “third party conflict management” refers to how that leader reacts to conflicts between two or more employees under his or her supervision. Past research on conflict management has predominantly focused on determining the best practices for managers, department leaders, or supervisors to intervene in employee conflict (e.g., Elangovan 1995; Nugent 2002). (In this article, we use “leader” to refer to all these functions.) For example, when the issue of conflict is considered important and a solution is urgently needed, it has been suggested that third parties should force a solution (Elangovan 1995; Nugent 2002).

What researchers have overlooked, however, is how employees' perceptions of leaders' conflict management behaviors may affect employees' well‐being (De Dreu and Beersma 2005). This question is important because recent studies have demonstrated associations between employee conflict and depression, declined self‐esteem, and decreased general health (De Raeve et al. 2009). Moreover, about 14 percent of people in Europe report work‐related health problems, such as stress, depression, and anxiety (European Union 2010). In turn, illnesses such as depression increase organizational costs because they are associated with absenteeism and decreased employee performance (Birnbaum et al. 2010). Workplace conflict can, therefore, have significant effect on organizational outcomes. Consequently, in the current study, we examined how employees' perceptions of leaders' conflict management behaviors affected the relationship between employees' experiences of workplace conflict and their levels of stress (see Figure One).

We conducted a field study to investigate the role of perceived leaders' third‐party conflict management behaviors on employees. With this study, we hope to contribute to the discussion of which factors affect the negative stressful impact of conflict on employee well‐being. Furthermore, we seek to develop additional insight into the effects of leadership behavior in conflict situations that can help design conflict interventions and conflict trainings for organizational leaders. For these purposes, we examined how employees perceived three of the most common types of conflict management behaviors (e.g., problem solving, forcing, and avoiding) displayed by leaders and how this affected their own conflict–stress relationship.

Past conflict research has distinguished between three types of conflicts among individuals in the workplace: relationship, task, and process conflict (Jehn 1995, 1997). Relationship conflict occurs when parties disagree about personal issues that are not work‐related, such as clashes of personality, political views, hobbies, and social events. Task conflicts occur when employees disagree about the task being performed, such as what is causing a work‐related problem and how they should solve it. Process conflicts are arguments about logistics (how to best achieve the agreed‐upon solution toa work problem) and delegation (how and to whom to delegate which tasks) (Jehn 1997).

In general, conflict in organizations can diminish parties' psychological well‐being (De Dreu and Weingart 2003; Jehn and Bendersky 2003). For example, imagine that an employee has outlined his or her opinion about how to solve a particular problem, but his or her colleague disagrees and argues that he or she is wrong. It is likely that the first person will experience some frustration and dissatisfaction. Past research has found that conflict increases negative emotions that, in turn, negatively affect individual well‐being by diminishing satisfaction and causing emotional exhaustion, which can increase absenteeism and employee turnover (Quick et al. 1997; Giebels and Janssen 2005). Workplace conflict may, therefore, have long‐lasting effects on individuals as well as organizations.

Conflict at work is a stressor (Keenan and Newton 1985), and all three types of conflict have been found to negatively affect employees' well‐being. Fights over task issues have been found to increase negative affect (Baron 1984), and to decrease satisfaction and intent to stay with the employer (Schweiger, Sandberg, and Ragan 1986). Previous research suggested that process conflict can have a negative impact on people's emotions (Greer and Jehn 2007; cf. Jehn and Bendersky 2003) and can increase the likelihood that a person will experience conflict in future interactions (Greer, Jehn, and Mannix 2008).

Both task and process conflict are associated with decreased well‐being, but to a lesser extent than relationship conflict (De Dreu and Weingart 2003; Behfar and Thompson 2007; Greer and Jehn 2007). Past research suggests that relationship conflict seems to have an even more detrimental effect on individual well‐being (compared with task or process conflict) because it can threaten one's identity and self‐esteem, and generate more intense emotions (De Dreu, Van Dierendonck, and Dijkstra 2004). Relationship conflict negatively affects morale, which is likely to result in decreased satisfaction with the job, group, and organization (cf. Jehn and Bendersky 2003). Furthermore, research has suggested that the different types of conflicts are often related to each other (Simons and Peterson 2000; Rispens 2012), indicating that people may misinterpret what the conflict is about. For example, one could perceive regular and ongoing conflict with colleagues about a particular task as a personal attack rather than a task‐related disagreement. Nevertheless, given the empirical evidence, any interpretation of the conflict — whether one sees it as a task, relationship, or process issue — is likely to negatively influence one's well‐being.

Organizational leaders typically fulfill an informal or emergent third‐party role in employee conflict (Pinkley et al. 1995; Kressel 2006). Leaders are usually involved parties, given their responsibility for constructive teamwork. They usually have a relationship with the conflicting parties beyond the conflict (e.g., Pinkley et al. 1995). Despite the fact that employees often believe that dealing with the conflict is one of the organizational leader's' responsibilities (Epitropaki and Martin 2004), a formal prescription of this role is often missing, and empirical research on organizational leaders as third parties has been rare (Goldman et al. 2008).

The third‐party role of organizational leaders differs from the role of institutional third parties who are external to and neutral in conflicts (e.g., outside mediators, institutional ombudsmen, etc.) in three ways (e.g., Pinkley et al. 1995). First, organizational leaders' performance heavily depends on their employees' performance. Second, organizational leaders often have an enduring relationship with their employees. Finally, an organizational leader may have his or her own interests regarding a specific outcome of the conflict between his or her employees (Lewicki and Sheppard 1985; Pinkley et al. 1995).

The psychoanalyst Karen Horney (1945, 1950) described the basic behavioral tendencies of people when faced with conflict. These three tendencies are: moving toward others, moving against others, and moving away from others. Past research on third‐party conflict has identified similar categories. For example, intravention is a combination of problem solving (moving toward) and forcing (moving against) behavior (Conlon, Carnevale, and Murnighan 1994). Others have found that leaders, as third parties, use autocratic behavior (moving against) to impose a settlement between the conflicting parties, or mediational behavior (moving toward) in order to gain insight into the conflicting parties' concerns and to stimulate them to find a solution themselves (e.g., Karambayya, Brett, and Lytle 1992). In addition, although this has been discussed less frequently, leaders confronted with conflicts may feel threatened, and therefore may try to avoid getting involved in the conflict (Sheppard 1983). Because of this, Robin Pinkley and her colleagues (1995) added leaders' avoiding behavior (moving away) to their dimensions of leaders' third‐party conflict management behavior.

In the current study, we examine the moderating effects of three corresponding third‐party behaviors of leaders — problem solving, forcing, and avoiding — on interpersonal conflicts between their employees. “Problem solving” is defined as searching for the underlying concerns of the parties and seeking to come to a solution that addresses all parties' concerns. “Forcing” occurs when the leader imposes on the disputants the solution that he or she prefers, or pushes for any resolution that will end the dispute. “Avoiding” occurs when the leader chooses not to get involved in the conflict.

Conflicts are likely to increase employee stress because they reduce employees' self‐esteem and diminish their sense of control over their situation (e.g., De Dreu, Van Dierendonck, and De Best‐Waldhober 2002). Leaders' problem‐solving behavior can involve asking the conflicting parties questions about their goals and points of view (e.g., Carnevale 1986); employees are likely to interpret this positively as the leader showing concern for their interests (cf. Giebels and Yang 2009). Hence, when employees perceive that they are allowed to express their viewpoints, their feeling of control over the conflict situation is likely to increase, and therefore can help buffer the conflict's stressful impact. Indeed, in their study, James Quick and Jonathan Quick (1984) indicated that a participatory leadership style, in which employees participate in the decision‐making process, decreases employees' feelings of stress.

Similarly, Ellen Giebels and Onne Janssen (2005) found that when outside help was called in, parties in conflict experienced fewer negative consequences in terms of individual well‐being than people who did not ask for third‐party help. Furthermore, a recent study by Renée de Reuver and Marianne Van Woerkom (2010) showed a negative correlation between leaders' engagement in problem solving with employees with whom they had conflicts and employee stress. Problem‐solving behaviors may have these effects because they demonstrate that the leader has a higher commitment to his or her employees, they increase the employees' perception of justice, and they enhance their sense that they have a voice in their workplace (de Reuver and Van Woerkom 2010). Although de Reuver and Van Woerkom (2010) focused on behavior in leader–employee conflicts, similar effects may exist for leaders' third‐party problem‐solving behavior. Also, William Ross, Donald Conlon, and Allan Lind (1990) suggested that the attention paid by third parties to people in conflicts (person‐oriented behavior) is important for maintaining feelings of satisfaction and fairness.

To summarize, the literature indicates that when conflicting parties perceive that their leader is engaging in problem‐solving behavior, they are likely to feel that their concerns are taken seriously, and consequently they experience less stress (Rhoades and Eisenberger 2002). Thus, our first hypothesis reads:

In contrast, when employees perceive that their leader is using forcing behavior, it is likely that the correlation between conflict and stress will be amplified. Forcing behavior is likely to increase the employee's feeling that he or she is losing control (Dijkstra, Van Dierendonck, and Evers 2005). Past research has shown that third‐party forcing behavior is negatively associated with perceptions of procedural fairness as well as with perceptions of the perceived fairness of the third party (e.g., Karambayya, Brett, and Lytle 1992). Injustice perceptions are stressors and are negatively related to psychological health (Judge and Colquitt 2004).

Moreover, forcing behavior is likely to be based on the third party's own interests, which may not be in line with the interests of the conflicting employees (Conlon, Carnevale, and Murnighan 1994). As a consequence, disputants' satisfaction with the conflict process or with decisions resulting from the forcing behavior may be low (Karambayya and Brett 1989), and conflict stress will increase. Using power directly to solve conflicts, without paying any attention to the underlying issues of concern, is unlikely to offer an ultimate solution to the situation (Peterson and Harvey 2009). Consequently, the conflict is likely to endure and may even intensify over time, accompanied with increased stress. Moreover, the leader's forcing behavior may only serve to pull the leader into the conflict (Peterson and Harvey 2009). This additional conflict can cause extra stress. We, therefore, come to our second hypothesis:

Making an effort to avoid a conflict situation is not a behavior consistent with the prototypical role that employees expect their leader to fulfill (Epitropaki and Martin 2004). Employees expect their leaders to have the authority and the obligation to settle conflicts among employees. Not doing so could be interpreted by the conflicting parties as a lack of support (Rubin, Pruitt, and Kim 1994; Hardy and Clegg 1996). When leaders avoid employee conflict, employees are likely to feel confused because they expected a different type of response.

Furthermore, the leader's avoiding strategy is likely to cause employee frustration. For example, when employees argue about which project must be cancelled because of a budget shortfall, they are likely to find it frustrating when they perceive their leader is avoiding the issue. When a leader fails to manage the conflict, the conflict may escalate, and the conflict–stress relationship may be intensified (Dijkstra et al. 2009). People's ability to process information decreases when they are experiencing conflict; consequently, they are less likely to change their opinion and consequent behavior. As such, the conflict could intensify and escalate. With escalating conflict, stress is likely to increase. Therefore, we propose that:

We invited all 341 employees of an insurance company to complete an online questionnaire. A total of 175 employees completed the questionnaire, for a response rate of 51 percent. Twenty‐four of the questionnaires were incomplete and removed from the data set.

Of the remaining 145 participants, 63 percent are female. The average age of respondents was 35.4 years (standard deviation = 8.1), and the average tenure with the organization was 6.6 years (standard deviation = 7.1). Sixty‐eight percent of the respondents had completed intermediate‐ or lower‐level vocational training, and 29 percent held a bachelor's or master's degree. The participating employees came from all departments of the company, such as claims and loss handling, human resources, call center, and marketing. The departments had an average size of 15.3 members (standard deviation = 12.9).

We measured relationship conflict and task conflict based on Karen Jehn's (1995) conflict scales. Process conflict was measured on the scale developed by Jehn and Elizabeth Mannix (2001). We measured each conflict type using a four‐item scale (see Table One for the specific statements). We asked participants to respond to these statements on a 7‐point Likert scale from (1) “almost never” to (7) “almost always.” We calculated the Cronbach's alpha's (measure for internal consistency) for relationship, task, and process conflict to be 0.86, 0.87, and 0.84, respectively. Cronbach's alpha's above 0.70 are expected to reflect internal consistency, meaning that the four items we used to measure a type of conflict were, indeed, reliable in the sense that they measured the specified construct (i.e., relationship, task, or process conflict).

Conflict stress was measured with four items developed by Ellen Giebels and Onne Janssen (2005) (see Table One for all items). Answers were measured on a 7‐point Likert scale from (1) “never” to (7) “always.” The Cronbach's alpha was 0.86. A confirmatory factor analysis with varimax rotation (see Table One) on the four conflict measures revealed four factors, each with an eigenvalue higher than 1. This indicates that the three types of conflict are distinct constructs and that they are also conceptually different from conflict stress. This is important because conflicts often involve tension and emotions, and could be confused with conflict stress.

We measured perceptions of the third‐party behaviors (forcing, problem solving, and avoiding) adapting items from the Dutch test for Conflict Handling (DUTCH) (Van de Vliert 1997; De Dreu et al. 2001). We rewrote the subscales to fit the third‐party role of leaders and asked employees to rate their direct leader or supervisor on these three behaviors. We measured the three different types of third‐party behavior using four statements (forcing and avoiding) and three statements (problem solving), each on a 5‐point‐Likert scale from (1) “completely disagree” to (5) “completely agree.” Cronbach's alpha was 0.72, 0.66, and 0.82 for problem solving, forcing and avoiding, respectively. A factor analysis with oblimin rotation, indeed, revealed three factors with an eigenvalue higher than 1 (see Table Two for all items and factor loadings).

Because the impact of leadership behavior may depend on the actual need for leadership (Hunter, Bedell‐Avers, and Mumford 2007), we controlled for that variable, measuring it with a three‐statement scale (Martin 1983). An example is “Results of my work performance would be better when there would be more leadership.” Answers were rated on a 7‐point Likert scale, from (1) “not at all” to (7) “to a high extent.” In addition, we controlled for gender, age, amount of hours worked per week, and department size (Siu et al. 2001; Matud 2004; Dijkstra, Van Dierendonck, and Evers 2005; Giebels and Janssen 2005).

To test our hypotheses, we conducted a series of hierarchical regression analyses. We ran these analyses separately for the impact of each type of conflict (relationship, task, and process) on conflict stress. To test our hypotheses about the effect of the leader's behavior (e.g., will the conflict‐related stress be higher or lower with different leadership behavior?), each analysis had several steps (Aiken and West 1991). In the first step, we entered the control variables. In the second step, we analyzed the predictor variables of conflict type and leader's third‐party conflict behavior to examine whether a main effect existed. In the third step, we added the interaction terms to reveal possible effects.

To minimize problems of multicollinearity and facilitate interpretation, we standardized the predictor variables before calculating the interaction terms and regression statistics (Aiken and West 1991). One could argue that given the fact that our respondents worked together in departments, our data could be nested, and therefore multilevel analysis should be used to test our hypotheses. However, calculation of the intraclass correlation (ICC‐1) values for our constructs indicated that our constructs did not have sufficient homogeneity within departments. Typical ICC‐1 values range between 0.05 and 0.20 (Bliese 2000). In our sample, the values were much smaller (between 0.00 and 0.02) with nonsignificant F‐values, which indicates that the variation between the departments and its leaders was not significantly higher than the variation within the departments. Accordingly, the differences between the departments were small, and the analyses could be done on the level of the individual employee instead of on a department level.

To control for the risk of multicollinearity (Cohen et al. 2003), we tested the variance inflation factor (VIF) and tolerance of all predictors. The VIF of the six predictors varied between 1.00 and 1.39; the tolerance of the six predictors varied between 0.72 and 1.00. Average VIFs close to 1.00 have little risk of multicollinearity (Bowerman and O'Connell 1990). Values of tolerance below 0.2 are reasons for concern (Menard 1995). Thus, multicollinearity was not a concern in our data, which means that the regression coefficient could be interpreted without high risks of misinterpretation. To interpret interaction effects, we conducted regression equations on conflict stress given conditional values for the predictors (M + 1SD; M − 1SD) (cf. Aiken and West 1991).

Because employees reported the different types of conflict as well as the dependent of variable conflict stress, we conducted the Harman's one‐factor test to examine the possibility of method bias. A principal component analysis on the three types of conflict, conflict stress, and perceived leader behavior failed to show one single factor or one general factor, indicating that overlap between different variables was small (Podsakoff and Organ 1986).

The correlations, means, and standard deviations of all constructs are listed in Table Three. Correlation analyses showed that the intercorrelations of the three types of conflict are significant and similar in value to those found in other studies (e.g., Simons and Peterson 2000; De Dreu and Weingart 2003). In Table Four, the regression coefficients of the control variables and the three main effects of type of conflict on conflict stress are shown.

In Table Four, the regression analyses are shown to test Hypotheses 1a, 1b, and 1c. Hypothesis One stated that relationship conflict (1a), task conflict (1b), and process conflict (1c) are positively correlated with conflict stress, and this relationship is affected by leaders' problem‐solving behavior such that the correlation between experiencing conflict and experiencing stress is stronger when leaders employ minimal problem‐solving behavior.

Results indicated that relationship, task, and process conflict all have significant and positive main effects on conflict stress, meaning that all three types of conflict are positively correlated with conflict stress. This result is consistent with our hypotheses. But the proposed impact of leaders' problem‐solving behavior was only significant in the case of relationship conflict. Our data suggest that when employees perceive their leader is engaging in problem‐solving behavior, their relationship conflict is less likely to cause them to experience stress (Figure Two). Therefore, we found support for Hypothesis 1a.

Hypothesis Two stated that relationship conflict (2a), task conflict (2b), and process conflict (2c) are positively correlated with conflict stress, and that this relationship is affected by leaders' forcing behavior such that the employees will report experiencing more stress when leaders employ a high level of forcing behavior. The results (see Table Five) revealed forcing behavior had significant impact on stress for each of the three types of conflict. Our data suggest that when employees perceive their leader is engaging in forcing behavior, their stress increases for all the three types of conflict situations (Figure Three). Hypotheses 2a, 2b, and 2c are, therefore, supported.

Hypothesis Three stated that relationship conflict (3a), task conflict (3b), and process conflict (3c) are positively correlated with conflict stress, and that this relationship is affected by leaders' avoiding behavior such that employees' stress levels increase when leaders employ avoiding behavior (Hypothesis Three). The results (see Table Six) showed a marginally significant effect on stress in instances of task conflict that were accompanied by leader avoidance behavior. Our data suggest that when employees perceive that the leader is engaging in avoiding behavior, task conflict in particular will cause them to feel more stress (see Figure Four). Hypothesis 3b is, therefore, supported.

The goal of this study was to investigate the impact of perceived leaders' third‐party conflict management behaviors on the relationship between employees' conflict and their levels of stress, which can have important impacts on their well‐being. Past research on conflict and employee well‐being largely ignored the role of the behaviors of employees' organizational leaders. This is surprising, considering the crucial role leaders play (Yukl 2000).

In a study of 145 employees of a Dutch insurance company, we found that leaders' conflict management behavior, as perceived by their employees, can have amplifying as well as buffering effects on the relationship between conflict and stress, depending on the type of conflict management behavior displayed. These results highlight how important it can be for informal third parties, such as organizational leaders, to deal with organizational conflicts to prevent them from diminishing employees' well‐being and subsequently damaging organizational functioning.

More specifically, conflict management behavior characterized as forcing was found to increase employees' stress experience for all three kinds of conflict (task, relationship, and process). A conflict‐avoiding leader, however, only amplified employees' stress when the conflicts in question were task‐oriented. Leaders' problem‐solving behavior decreased employees' stress levels when the conflicts were relationship‐oriented. Thus, we suggest that conflict researchers should examine more thoroughly the behaviors of leaders, how employees perceive leaders' third‐party conflict behavior, and the impact of this behavior on employees' health and well‐being.

An important finding is that the problem‐solving behavior of leaders can buffer the detrimental effects of relationship conflict on individual well‐being. This is interesting because relationship conflicts are thought to be more difficult to resolve when compared with task or process conflicts (De Dreu and Van Vianen 2001). Problem‐solving behavior includes asking each conflicting party about his or her point of view (e.g., Carnevale 1986), which is likely to be interpreted by each conflicting party as paying attention to his or her interests (cf. Giebels and Yang 2009). Apparently, the fact that a leader inquires about each party's viewpoints and feelings enhances the employee's feeling of control over his or her situation. When a leader is willing to listen to conflicting employee opinions and emotions, he or she demonstrates concern for employees' well‐being (e.g., Lyons and Schneider 2009).

It also seems likely that when a leader deals with employee relationship conflicts by using problem‐solving techniques, the employees perceive that the leader seeks to create some common orientation or common ground (i.e., behaving professionally) despite the interpersonal differences among her staff.

Another possible explanation, however, is the misattribution of task conflict as a relationship conflict (Simons and Peterson 2000; Rispens 2012). Tony Simons and Randall Peterson (2000) found that the correlation between task and relationship conflict was lower in teams with a high level of trust than in teams with a low level of trust. When leaders manage task conflict in a problem‐solving way, the personal animosity among the conflicting parties that may have caused them to perceive personal attacks (i.e., relationship conflict) could be blocked. Problem‐solving leaders are likely to listen to each party's point of view and to encourage understanding between the parties. Emotional and personal issues are likely to be addressed, which allows greater focus on the task issues. The decreased emotional involvement may help the parties discuss the task and/or process issues in a more productive manner. As mentioned earlier, task and process conflict may cause stress as well; however, to a lesser extent than relationship conflict, accordingly employees still experience some kind of stress.

This study extends the discussion of how managing relationship conflict can decrease its negative effects on employees. These results and the explanations for the positive effects of problem‐solving behavior appear similar to the explanations that have been offered to explain the success of mediation and its impact on well‐being (e.g., satisfaction, justice, and agreement; see for a review Wall, Stark, and Standifer 2001 and Wall and Dunne 2012).

Furthermore, our results point toward the utility of differentiating among task‐, process‐, and relationship‐related issues in determining the effects of leader behavior. An explanation for the differences between results according to types of conflict in our results could involve employees' expectations of a prototypical leader (Epitropaki and Martin 2004). A prototypical leader is responsible for his or her employees and their tasks. In this way, task‐related and process‐related conflict issues could be seen as the leaders' responsibility to solve. Accordingly, employees would be likely to perceive that engaging in problem‐solving behavior in task and process conflicts is their leader's duty more than they would when the conflict is more personal or relationship‐oriented. This expectation may mean that the leader's involvement in task and process conflicts would have less impact on their experience of stress. This explanation is supported by the finding that the leader's avoiding (e.g., passive) behavior amplifies the stress, especially in task conflict. In task disagreements, employees' expectation that their leader will help solve the problem may be high — when leaders do not act according to their expectations, employees may become disappointed and frustrated. Future research should focus on the underlying mechanisms of employees' expectations and the needs for leaders to better understand why some of their behaviors are effective and others are not.

Our study confirms that the direct expression of power in the form of forcing behavior can harm employees' well‐being (cf., Peterson and Harvey 2009). A forcing leader may become an additional party to the conflict (i.e., employees may turn against their leader). This creates an even more complex situation for employees and can increase tension and negative emotions. Alternatively, the involvement of a leader in a conflict between employees may be perceived by the conflicting parties as an indication that they failed to effectively deal with the issue themselves, and therefore “lost face” (Ting‐Toomey and Kurogi 1998). In addition, because leaders judge employee functioning and performance, the conflicting parties may perceive that the leader's forcing behavior indicates that they do not function well, increasing their anxiety. Nevertheless, sometimes leaders may feel it necessary to use forcing behavior, for example, when time is limited and the need for a solution is significant (Nugent 2002). In weighing whether or not to use forcing behaviors to address employee conflict, leaders should be aware of the detrimental effects this behavior can have on employees' well‐being.

Our results also shed light on the differential impacts of the different types of conflict responses. For example, we found that problem solving affected only relationship conflict, and avoiding behaviors affected only task conflict. This highlights the assumption that different types of conflict should be managed differently.

Our study has implications for organizational leaders who seek to manage employee conflict and for organizations who seek to reduce the detrimental consequences that employee conflicts can have on their staff and their organizations. Organizational leaders should be trained to recognize the different types of conflict and how to manage them, with a focus on the impact of conflict on employee well‐being. Because our research focused on perceptions of leader behavior, they suggest that leaders should be particularly aware of how employees perceive their behavior. Accordingly, leaders should check to see how their problem‐solving intentions are perceived. Moreover, leaders should clarify their behaviors and intentions so that they are not seen as conflict‐avoidant.

In this study, conflict stress was measured as an outcome of interpersonal conflict. One may argue that conflict‐related stress is a short‐term consequence of conflict, and can therefore be disputed as a significant outcome variable. Earlier studies, however, have found long‐lasting effects of stress on individual health (e.g., Reznik, Roloff, and Miller 2010) and support the idea that stress has a significant impact on individuals and organizations. Giebels and Janssen (2005) found strong relationships between conflict stress and three indicators of individual well‐being (absenteeism, turnover intentions, and emotional exhaustion) that have important impacts on organizations in the long run, highlighting the relevance of conflict‐related stress. To understand the role of leader's conflict management and its effects more deeply, however, we recommend future research to examine the effects of leader's third‐party conflict management on other outcomes (e.g., performance, productivity, decision quality, and innovative behavior).

This study measured employees' perceptions of leaders' third‐party behavior, rather than the actual leader behavior. We must, therefore, acknowledge that the perception of the leaders' behavior may differ from their actual or intended behavior. Therefore, we suggest future research could include controlled experiments or observational studies to examine the leaders' actual behavior.

Furthermore, our results should be cautiously interpreted because of the cross‐sectional design of the study. We are not able to test whether conflict stress is a consequence of conflict or an antecedent. However, the statements we used to measure conflict stress were explicitly related to conflict (e.g., “after a conflict I feel upset”) in ways that make an alternative explanation less likely.

Future research should use objective data, such as absenteeism and employee performance, to verify our results on the subjective measure of well‐being (e.g., perceived stress). To further contextualize our understanding of the role of leadership behavior, future research could take cultural aspects into account. For example, relative power distance (Hofstede, Hofstede, and Minkov 2010) could influence how employees perceive leader interventions. Power distance is relatively low in the Netherlands, and forcing behavior may be perceived differently in cultures with higher power distance because people in such cultures are more likely to accept behavior that the Dutch might perceive as authoritarian.

Gender could also be a factor — other studies have found that the third‐party interventions of women yield different results than those of men (Benharda, Brett, and Lempereur 2010). Finally, future research could examine such additional leader conflict behaviors as yielding and compromising (Van de Vliert 1997).

We suggest that an employee's perceptions of how a leader has behaved as a third party to a conflict can amplify as well as buffer the employee's experience of stress due to workplace conflict. These are important findings because dealing with conflicts is a major task of organizational leaders. Based on our results, leaders should be aware of the effects their behavior can have on employees' conflict‐related stress. Specifically, forcing and avoiding behavior need to be used cautiously. Problem‐solving behavior, particularly in relationship conflict, can help alleviate the stress experienced by colleagues in conflict.