SEARCH

SEARCH BY CITATION

Keywords:

  • task conflict;
  • relationship conflict;
  • intragroup trust;
  • competing hypotheses testing

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

Although previous research have reported strong and positive intercorrelations between group task conflict and group relationship conflict, several different theoretical rationales exist for the positive link and the empirical literature remains equivocal. To clarify the causal linkage between the two types of conflict, we derived seven models specifying how group task and relationship conflict can be related to each other, including some rival explanations. We tested the competing models using a longitudinal panel design, with data collected from 74 project teams comprising a total of 388 students. The results indicated that relationship conflict led to an increased subsequent task conflict through negative group affect. Task conflict, however, predicted a subsequent relationship conflict under a specific context, that is, groups that had lower levels of trust among the members. Copyright © 2010 John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

Beyond past debates over whether conflict is bad or not, several authors have advanced to the point of what type of conflict is detrimental or beneficial to groups (e.g., Amason, 1996; Guetzkow & Gyr, 1954; Jehn, 1994). One prominent typology of conflict has been made as to whether it is task-related/substantive (task conflict) or socio-emotional/interpersonal (relationship conflict). Task conflict is primarily related to performing tasks, which is often proposed to improve the quality of group work by encouraging more alternative ideas and to help a group avoid “conformity traps” (Torrance, 1957) or “groupthink” (Janis, 1972). Relationship conflict, on the other hand, refers to conflict caused by interpersonal animosity. Relationship conflict causes interpersonal hostility and disturbs harmony; thus, it is regarded as detrimental to groups (Amason, 1996; Jehn, 1995).

Early studies on task and relationship conflict attempted to demonstrate the distinct effects of each conflict type on group performance: positive effects of task conflict and negative effects of relationship conflict (e.g., Amason, 1996; Jehn, 1994). However, a recent meta-analytic review by De Dreu and Weingart (2003) suggested that it is not only relationship conflict that is detrimental to a group but task conflict as well.

There have been debates among conflict theorists on whether or not both types of conflict are harmful. Although scholars have reached a consensus with respect to the detrimental effects of relationship conflict, this is not the case with task conflict (Spector, 2008). De Dreu (2008), in his review of related literature, argued that the benefits of task conflict are constrained to a very narrow set of situations and that the costs outweigh the benefits. Jehn and Bendersky (2003) delved into situations that could possibly moderate the effects of conflicts on team outcome, and suggested that such variables as task interdependency, task routineness, group norms, conflict management processes, and emotions can amplify, ameliorate, exacerbate, or suppress the relationship between them. For example, Van de Ven and Ferry (1980) suggested that task conflict benefits groups under non-routine task situation, whereas its effects are dysfunctional under routine task situation as it interferes efficient group processing. Tjosvold (2008) argued that constructive controversy, a well-managed conflict, can lead to a greater understanding of organizational issues and quality decision-making in a wide range of situations.

Whether task conflict can be functional or not, one known big drawback of task conflict is that it is tightly connected with relationship conflict (De Dreu & Weingart, 2003; Simons & Peterson, 2000). De Dreu and Weingart (2003) found an average corrected correlation of 0.54 between the two constructs in their meta-analytic review.

However, the issue of why relationship conflict and task conflict are related to each other has not been fully investigated. Although some researchers directly draw their attention to the issue (Peterson & Behfar, 2003; Simons & Peterson, 2000; Yang & Mossholder, 2004), their research focus was limited almost exclusively to task conflict triggering relationship conflict, or they could not fully investigate the origin of the intercorrelation because of their cross-sectional designs. This raises the following research questions that the present research aims to address: Why and how are the two variables related? Can either of them be a cause of the other, or are they reciprocally related? Do any conditions or circumstances change the direction and/or strength of the association? Drawing from relevant psychology and management theories, we pose seven possible models specifying the associations between task and relationship conflict. Among the proposed models, some present rival explanations for the nature and direction of causality between the two. We test the proposed models using a longitudinal panel design.

Models of Task and Relationship Conflict

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

A graphical depiction of the proposed models is provided in Figure 1. Although some of the models have received more attention than the others, each provides a unique perspective on the associations between the two types of conflict. Nonetheless, they appear to be, at least to a certain extent, mutually exclusive in that it is unlikely that all could hold.

thumbnail image

Figure 1. Models of the possible associations between task and relationship conflict

Download figure to PowerPoint

In the following sections, we present a rationale for each model and develop several alternative hypotheses on the causal link between task and relationship conflict. To test alternative hypotheses simultaneously, we adopted a competing hypothesis approach. It was suggested that when two or more reasonable explanations can be derived from prior knowledge, the competing hypothesis approach is more appropriate than the exploratory approach (no formally stated hypothesis) or the dominant hypothesis approach (a single hypothesis; see Armstrong, Brodie, & Parsons, 2001).

Direct causal linkage

Early studies on conflicts in organizational context assumed that task and relationship conflicts are mutually independent and each conflict type brings different effects on group performance. They investigated the effects of each conflict type on outcome variables separately; thus the issue of connectedness between the two types of conflict was not highlighted. However, most studies on task and relationship conflicts manifested a positive correlation between them (see De Dreu & Weingart, 2003; Simons & Peterson, 2000 for review). Recently, several scholars have begun to turn their attention to the issues of why and how task and relationship conflict are related to each other. In the subsequent models below, we explored possible explanations for the connectedness between the two types of conflict. Models 1–3 are presented to examine possible direct causal relationships between the two types of conflict: (1) a causal link from task conflict to relationship conflict (Model 1), (2) a causal link from relationship conflict to task conflict (Model 2), and (3) a reciprocal causal effect between the two (Model 3).

Model 1: Task conflict leads to relationship conflict

Model 1 posits that task conflict leads to relationship conflict. Ross (1989) argued that task-related disagreements are not always purely substantive but often contain emotionally harsh language that can be taken personally. Torrance (1957) also stated that disagreement itself can be perceived as evidence of personal rejection. Research on emotional misattribution effects supports this view. For example, West (1975) found that people who confronted a counter argument directly toward them made more extreme attributions regarding their partners' attitudes and became more personal than those people who confronted a counter argument toward people in general. In line with this view, several scholars have argued that task conflict may transform into relationship conflict in the group level as well (Jehn, 1997; Pelled, Eisenhardt, & Xin, 1999; Ross, 1989; Simons & Peterson, 2000). For example, group members whose ideas on a particular issue are disputed or rejected by other members may assume their competence is being challenged (Tjosvold, 1992), feel irritated, and attribute this task-related argument as a personal attack (Jehn, 1997; Pelled et al., 1999). Empirical evidences that support the causal linkage from task conflict to relationship conflict are provided by Mooney, Holahan, and Amason (2007), Peterson and Behfar (2003), and Simons and Peterson (2000).

Model 2: Relationship conflict leads to task conflict

In contrast to Model 1, the opposite prediction may be possible: Relationship conflict causes task conflict. Although a few scholars have suggested that relationship conflict may trigger task conflict (Jehn, 1995; Pelled et al., 1999), this causal sequence has received little empirical attention. Existing research that articulates the role of distorted perception in social conflict (see Cooper & Fazio, 1979; Pruitt & Rubin, 1986, for reviews of this body of research) provides a rationale for Model 2, that is, why and how negative attitudes toward disliked group members are formed, sustained, and further transformed into disputes and controversies among group members. In particular, a negative halo effect bias may serve as a psychological mechanism in the process of translating relationship conflict to task conflict.

A halo effect is a tendency to influence the judgment one makes about another individual by relying on a global impression of the individual (Thorndike, 1920). Research on halo effect biases demonstrates that people tend to give a favorable statement about another person who is regarded positively, whereas an unfavorable statement is given to a person who is regarded negatively (Goldman, Cowles, & Florez, 1983; Kelley, 1950). In particular, a negative halo effect may explain how relationship conflict is transformed to task conflict. That is, a member will tend to form a negative impression toward other members as a result of personal conflict and to disagree with or reject their ideas. Evidence was provided by Eisenhardt and Bourgeois (1988) who found that executives in political infighting failed to engage in natural issue-based debates, but rather tended to find fault with others' ideas.

Model 3: Reciprocal association between task and relationship conflict

Model 3 postulates a reciprocal association between task and relationship conflict. Although the co-occurrence of both types of conflict has been mentioned (e.g., Jehn, 1997; Ross, 1989; Simons & Peterson, 2000), little previous research has formally tested a mutual causal effects model. In the present study, Model 3 presents a hybrid model of Models 1 and 2. If both models are supported, it suggests that each conflict type may be caused by or generate the other type of conflict.

Models 1, 2, and 3 commonly assume a direct linkage between task and relationship conflict, but the directions of causality differ with each sound theoretical rationale. Instead of arguing one dominant hypothesis, we set three competing hypotheses and test the plausibility of each model empirically. From Models 1, 2, and 3, we derived three competing hypotheses as follows:

  • Hypothesis 1a: Task conflict will increase relationship conflict in groups.

  • Hypothesis 1b: Relationship conflict will increase task conflict in groups.

  • Hypothesis 1c: Task conflict and relationship conflict in groups will be reciprocally related to each other. In other words, group task conflict will predict, and be predicted by group relationship conflict.

Mediation linkage

Most evidence showed a strong correlation between task and relationship conflicts, but theoretical mechanisms underlying the relationship between the two have been less clear. To elaborate on the direct effect models, we were interested in exploring intervening mechanisms that might underlie the relationships. Meditational models explain the relationship between the two types of conflicts, as well as specify more precisely the sequence of effects that lead to task or relationship conflict. This study highlights the role of group-level affect (Barsade, 2002; Bartel & Saavedra, 2000) in linking the two conflict types.

Several scholars suggested that affective experiences are strongly related to conflicts. For example, Jones (2000) argued that conflict does not exist in the absence of emotion. Lindner (2006) attended to the interplay of emotion and conflict: emotion affects conflict, and conflict in turn influences emotion. At the group level, research also suggested that the negative affective experiences by group members can be both an antecedent and a consequence of conflict (Barsade, 2002; Bartel & Saavedra, 2000). Although these scholars did not distinguish the dimensions of task and relationship conflicts, we extend their works and examine the role of group-level affect in transforming one conflict type into the other. We particularly focus on the centrality of the negative group affect in conflict transformation processes and present two possible models to empirically examine the mediating effects of negative group affect on the associations between task and relationship conflicts.

Model 4: Negative group affect mediates the task conflict → relationship conflict linkage

Recent research on affect claims that group members can develop a mutually shared affect via the same processes operating in dyads, such as emotional contagion and comparison processes (Barsade, 2002; Bartel & Saavedra, 2000; George, 1990; Totterdell, 2000). How a group member feels is apt to be perceived by other group members during verbal and nonverbal conscious and non-conscious interactions, which then elicit similar affective reactions from others, resulting in some degree of convergence of group members' affect (Totterdell, 2000).

Affect is used as a general term for mood, emotion, and dispositional affect (Weiss & Cropanzano, 1996). Scholars distinguished mood from emotion (Barsade, 2002; George & Brief, 1992). Emotions are intense and relatively short-term affective reactions to a particular event, whereas moods are weaker, more diffused, and relatively enduring affective reactions than emotions. Scholars who characterized affect as a group-level phenomenon focused on moods to capture more of the day-to-day affective experiences of group members (Barsade, 2002; Bartel & Saavedra, 2000). In research, such terms as group affective tone (George, 1990; Varela, Burke, & Landis, 2008) and group affective climate (Gamero, González-Romá, & Peiró, 2008) are used as group affect constructs, and both are comparable concepts referring to a shared mood among group members (Pirola-Merlo, Hartel, Mann, & Hirst, 2002).

Model 4 proposes that negative group affect may mediate the effect of task conflict on relationship conflict. That is, task conflict is likely to evoke negative affective responses from group members, thereby creating hostilities.

Evidence for the link between task conflict and negative group affect can be found in several works. For example, Gero (1985) found that consensus groups (comparable to low task-conflict groups) were more likely to experience a friendly group climate than majority-rule groups (comparable to high task-conflict groups). Priem and Price (1991) reported a similar finding that group affect was friendlier in consensus groups than devil's advocacy groups (high task-conflict groups). Welty (1983) also suggested that a debate tends to generate a more combative environment than a consensus.

The effect of group negative affect on relationship conflict is evidenced by literature on the role of a negative (or positive) affective state in driving the subsequent behaviors. For example, negative affect leads to hostile interactions (Carnevale & Isen, 1986), while positive affect leads to interpersonal attractions (Gouaux, 1971), sociability, generosity, and helpfulness (Isen, 1970; Isen & Levin, 1972). Carnevale and Isen (1986) found that during the negotiation process, people in a positive state were less likely to use combative tactics and more likely to arrive at integrative solutions compared with people who were not in a positive state. At the group level, Barsade (2002) showed that positive emotional contagion among group members decreased the destructive mode of conflict.

Model 5: Negative group affect mediates the relationship conflict → task conflict linkage

Similar to Model 4, the direct effect of relationship conflict on task conflict (Model 2) might be more accurately specified through negative group affect. Specifically, relationship conflict among group members leads them to experience negative group affects, and therefore they are more likely to disagree with one another. Apart from controversy with another person, negative interpersonal relationship itself also has a high arousal potential which tends to generate a negative affective state (Berscheid, 1983). This negative affect, in turn, is expected to create more disagreements with another person. The mood congruent theory (Milberg & Clark, 1988) provides a theoretical explanation for our argument. Milberg and Clark (1988) found that people in a positive mood (e.g., happy) were more likely to think positively about others, which led them to comply or agree with the others. On the other hand, people in a negative mood (e.g., angry) tended to think negatively about others, which led them to disagree with the others. Although the above arguments are largely based on dyadic or individual-level research, it is expected that the same explanation can be applied to groups.

Based on Models 4 and 5, we propose the following competing hypotheses:

  • Hypothesis 2a: The association between task conflict and subsequent relationship conflict will be mediated by negative group affect.

  • Hypothesis 2b: The association between relationship conflict and subsequent task conflict will be mediated by negative group affect.

Moderation linkage

Recently, those scholars who focused on the high intercorrelation between the conflict types attempted to search for contextual factors that could reduce the risk of task conflict triggering relationship conflict: intragroup trust (Peterson & Behfar, 2003; Simons & Peterson, 2000; Tidd, McIntyre, & Friedman, 2004), social interaction (Gamero et al., 2008), emotionality, and intragroup relational ties (Yang & Mossholder, 2004). Among these moderators, intragroup trust has been studied most often and found to decrease the spillover of task conflict to relationship conflict (Peterson & Behfar, 2003; Simons & Peterson, 2000). There is, however, limited research examining the factors that moderate reverse causality. Thus, we propose that intragroup trust may reduce the spillover of relationship conflict to task conflict as well. The following models suggest the role of intragroup trust in moderating the task-relationship conflict linkage (Model 6) and the relationship-task conflict linkage (Model 7).

Model 6: Intragroup trust moderates the task conflict → relationship conflict linkage

Consistent with Simons and Peterson (2000), Model 6 suggests that intragroup trust lowers the effects of task conflict on relationship conflict: When group members show a higher level of trust in one another, task conflict is less related to relationship conflict because they are less likely to misattribute task-related disagreements as personal in motive or as personal attacks. On the other hand, when group members do not trust one another, task conflict is more closely and positively related to relationship conflict.

Model 7: Intragroup trust moderates the relationship conflict → task conflict linkage

In contrast to Model 6, intragroup trust may moderate the opposite direction. Researchers have devoted little attention to this reverse causality. Although Simons and Peterson (2000) pointed out the feasibility of the trust × relationship conflict interaction, they could not directly test it partly due to their cross-sectional data. We propose that intragroup trust may also alleviate the transformation of relationship conflict into task conflict. As proposed in Model 2, group members who develop a negative attitude towards others with whom they have relationship conflict are likely to judge others less favorably and tend to disagree with them. If this distorted perception on hostile members fosters task conflict within groups, then the degree of trust on the relationship of group members could play a moderating role. That is, when group members trust one another, they are less likely to disagree or refute disliked members' opinions by relying on their negative impression on the persons. When group members do not trust one another, relationship conflict is more likely to trigger task conflict.

On the basis of the above discussion, Hypotheses 3a and 3b are derived as follows:

  • Hypothesis 3a: The level of group trust will moderate the effects of task conflict on relationship conflict in groups, in such a way that task conflict generates higher levels of relationship conflict in groups with lower levels of intragroup trust, but not in groups with higher levels of intragroup trust.

  • Hypothesis 3b: The level of group trust will moderate the effects of relationship conflict on task conflict in groups, in such a way that relationship conflict generates higher levels of task conflict in groups with lower levels of intragroup trust, but not in groups with higher levels of intragroup trust.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

Participants and procedures

Participants for this study included 388 undergraduate students enrolled in eight introductory organizational behavior classes taught by the same instructor at a large university in the United States. Data collection was conducted for four consecutive semesters (two academic years). The students participated in this study for partial course credit and were randomly assigned to a total of 74 groups at the beginning of the semester. The groups were composed of four to seven members. Initially formed groups were retained until the end of the semester without changing their members. Over the 13-week term, group members had chances to interact frequently with one another by engaging in a group project. The group project was a case study based on theories or ideas taken from the course materials. Each group selected one local company, conducted interviews with its employees, analyzed its findings, and wrote a case on organizational behavior issues in the company. Thus, the group project required a great deal of collaboration for an extended period of time.

The study utilized a cross-lagged panel design with two waves, each involving an identical survey procedure and instruments. In addition, we measured negative group affect for the mediation analysis between the two waves. Thus, we collected data at three different points in time. We selected the time lag based on Gersick's (1988) punctuated equilibrium model because an appropriate time lag for the longitudinal relationships between the study variables is unknown. According to Gersick's findings, group members experience a period of inertia until the first half of their lives and then become more energized with their work in the second half. We expected that substantial changes in task and relationship conflict were most likely to occur in the second half of the group's existence. The first survey that measured the baseline levels of conflict was conducted before the mid-point, and the second and the third surveys were conducted after the mid-point.

The first survey was administrated on weeks five and six in class. By that time in the semester, the group members were well acquainted with one another and were about to focus their attention on the project's completion. At time one (T1), 378 (a response rate of 97.4 percent) usable responses were obtained after excluding patterned responses (e.g., three's checked for all items) and those returned beyond the time frame. The second survey was administrated on weeks 10 and 11. Only group affect was measured at time two (T2), with 365 (a response rate of 94.1 percent) responses deemed usable. The third survey took place on the last week of the semester, approximately seven weeks after the first data collection. At time three (T3), the same procedure and instruments were used as in T1. A total of 361 (a response rate of 93.0 percent) usable responses were obtained. Sixty-five percent of the participants were male. The mean age of the participants was 22. Sixty-two percent were Caucasian, 28.8 percent were Asian, 4.5 percent were African American, 3.8 percent were Hispanic, and three respondents did not report.

Measures

Task conflict and relationship conflict

We measured task conflict and relationship conflict at T1 and T3 using Jehn's (1995) task and relationship conflict questionnaire. The questionnaire consists of eight items (four items for each conflict type) with a five-point Likert scale. The wording of some of the original items was slightly modified for consistent phrasing. For example, “people in your work unit” in the original items was rephrased to “members of your team.” A sample item for task conflict is “how often do members of your team disagree about opinions regarding the work being done?” A sample item for relationship conflict is “how much friction is there among members of your team?” Coefficient alphas for these scales were 0.80 and 0.86 for task conflict T1 and T3, respectively; 0.86 and 0.91 for relationship conflict T1 and T3, respectively.

Intragroup trust

Intragroup trust was measured at T1 with the five items adapted from Simons and Peterson (2000). The response format consists of seven-point scale, anchored by one (never) and seven (always). A sample item for intragroup trust is “members of your team absolutely respect each other's competence.” The coefficient alpha for intragroup trust scale was 0.89.

Negative group affect

We measured group affect by aggregating self-reports of individual affect, which has shown to be valid and consistent with observers' ratings of group affect (Barsade, 2002; Bartel & Saavedra, 2000; Gamero et al., 2008). Individual affect was measured by asking each participant to assess the negative affect experienced in his or her group. We used a series of negative affect items adapted directly from Suls, Martin, and David's (1998) work. The items were “tensed,” “unhappy,” “irritated,” “angry/hostile,” “worried/anxious,” and “depressed/sad.” When rating the affect items, participants were instructed to use one-month response frame (e.g., during the past four weeks, not on the particular day). An underlying assumption on the use of this long-term time frame is that individuals' affect experienced in their groups may not be changed readily, but is relatively enduring as it is shaped through the course of mutual interactions (Bartel & Saavedra, 2000; Gamero et al., 2008). It is also noteworthy that negative group affect, a mediator, was measured at T2, separating it from its predictors at T1 and dependent variables at T3. This practice was recommended to minimize potential common method biases (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). A seven-point scale was used (one = definitely does not apply to your feelings to seven = definitely describes your feelings). The coefficient alpha for individual affect scale was 0.89.

Control variable

Group size, gender diversity, race diversity, and perceived value diversity were included as controls in the analyses. In previous research, these attributes have been considered as possible factors influencing both task and relationship conflict: group size (Amason & Sapienza, 1997), gender diversity (Jehn, Northcraft, & Neale, 1999), race diversity (Chatman, Polzer, Barsade, & Neale, 1998; Pelled et al., 1999), and perceived value diversity (Hobman & Bordia, 2006; Jehn, 1994; Jehn et al., 1999). Group size was computed as the number of members in a group. Based on demographic measures collected from the study survey, we computed gender diversity and race diversity using Blau's (1977) index of heterogeneity (1 − Σi2, where i is the proportion of respondents in the ith category). This index has been frequently used in previous studies (e.g., Kearney, Gebert, & Voelpel, 2009; Simons, Pelled, & Smith, 1999) to measure the degree of diversity when the variable is measured by categorical scales. Perceived value diversity was measured at T1 and T3 using Jehn et al.'s (1999) six-item measure. Participants responded to these items on a five-point scale, ranging from one (strongly agree) to five (strongly disagree). Sample items include, “The values of all members of your team are similar” and “Your team as a whole has similar work values.” The coefficient alphas for this scale were 0.82 at T1 and 0.91 at T3.

At the initial testing of the research hypotheses, the results with and without the control variables were essentially identical. Thus, we excluded control variables from our structural equation modeling (SEM) analyses for a more succinct presentation.

Confirmatory factor analysis

We conducted a confirmatory factor analysis to assess the factor structure of our measures. As suggested by Hu and Bentler (1999), we used several fit statistics to evaluate the overall model fit: the comparative fit index (CFI), the normed fit index (NFI), and the root mean square error of approximation (RMSEA). First, we assessed the convergent validity of eight study variables: task conflicts T1 and T3, relationship conflicts T1 and T3, intragroup trust T1, group negative affect T2, and value diversities T1 and T3. The fit statistics supported convergent validity for an eight-factor model, χ2 (660, N = 388) = 1270.19, CFI = 0.98; NFI = 0.97; RMSEA = 0.05. Item loadings on separate factors all exceeded 0.52 (p < 0.001). For a discriminant validity check, we conducted chi-square difference tests between the proposed eight-factor model and the two other rival models. We were particularly interested to see if respondents were able to distinguish task conflict from relationship conflict at both time periods. The eight-factor model fits the data significantly better than the seven-factor model, where the correlation between task and relationship conflict at T3 was constrained to 1.0, χ2 (661, N = 388) = 1336.15, Δχ2 (1, N = 388) = 65.97, p < 0.001, and the six-factor model in which the correlations between the two at both T1 and T3 were set at 1.0, χ2 (662, N = 388) = 1466.46, Δχ2 (2, N = 388) = 196.27, p < 0.001. These results support convergent and discriminant validity for our observed measures.

Aggregation

The research hypotheses of this study were tested at the group level. Thus, all individual responses were aggregated to the group level by taking the average score of the group. To justify aggregation, following the suggestion by Klein and Kozlowski (2000), we examined multiple aggregation indices: Eta-squared, rwg, and intraclass correlation coefficients (ICC) (1). First, eta-squared statistics for all variables were above the minimum value of 0.20 (Georgopolous, 1986): 0.33 and 0.39 for task conflicts T1 and T3, respectively; 0.33 and 0.43 for relationship conflicts T1 and T3, respectively; 0.31 for intragroup trust T1: 0.42 for negative group affect T2; 0.67 and 0.64 for value diversities T1 and T3, respectively. Second, the rwg (James, Demaree, & Wolf, 1984) statistics also supported the aggregation. All the average rwg values ranged from 0.85 to 0.91, exceeding the common threshold point of 0.70, which indicates relatively high levels of agreement within groups. Finally, ICC (1) for each variable was calculated to check if the individual group members' opinions were interchangeable within the group. The F-test for the ICC (1) value for each measure was significant (p < 0.01), indicating the appropriateness of the aggregation.

Analyses

We conducted SEM with maximum-likelihood estimation (AMOS 16; Arbuckle, 2007) to test the study hypotheses. For each construct, a single summed indicator was used to represent the latent variable. The use of single summed indicators substantially improved the ratio of the number of responses to the number of parameter estimates. Thus, we were able to incorporate the effects of measurement error in structural parameter estimation (Anderson & Williams, 1992; Williams & Hazer, 1986) with a relatively small sample size. Following the procedures by Williams and Hazer (1986), we set the values of the measurement parameters before testing the models. Specifically, the path from each construct to its indicator was fixed to the square root of the reliability for the measure, while the error variance for each indicator was fixed at (1 − reliability) × the variance of the indicator. In order to minimize potential biases caused by a third unmeasured variable, our analyses also included residual correlations by allowing correlations among residuals (see Anderson & Williams, 1992).

We checked the normal distribution of the data. The kurtosis values for all study variables were not significant (all kurtosis values < 1.1, ns). The Mahalanobis distance statistics suggested that there was no significant outlier in the sample. We also tested multicollinearity among a set of predictor variables by assessing the variance inflation factors (VIFs). The VIFs ranged between 1.1 and 2.1, which indicates that the intercorrelations among predictor variables are unlikely to be problematic (Nester, Kunter, Nachtsheim, & Wasserman, 1996).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

Table 1 presents descriptive statistics for the group level variables. As can be seen, task conflict and relationship conflict were significantly related to each other both cross-sectionally and longitudinally.

Table 1. Means, standard deviations, and correlations for study variables
VariableMSD1234567891011
  • Note: N = 74 groups. Reliabilities are in parentheses along the diagonal; T1, Time 1; T2, Time 2; T3, Time 3.

  • *

    p < 0.05;

  • **

    p < 0.01.

1. Group size5.270.75          
2. Gender diversity0.400.190.22         
3. Race diversity0.390.230.05−0.06        
4. Value diversity T12.320.370.01−0.110.23*(0.88)       
5. Value diversity T32.290.490.19−0.130.070.57**(0.95)      
6. Task conflict T12.190.360.10−0.140.070.50**0.45**(0.89)     
7. Task conflict T32.310.470.210.000.030.52**0.58**0.59**(0.92)    
8. Relationship conflict T11.670.390.200.000.130.53**0.53**0.61**0.61**(0.91)   
9. Relationship conflict T31.920.570.31**0.080.070.46**0.64**0.42**0.83**0.66**(0.95)  
10. Negative group affect T24.310.530.19−0.140.130.54**0.70**0.46**0.59**0.59**0.64**(0.89) 
11. Intragroup trust T15.200.59−0.200.12−0.08−0.69**−0.55**−0.37**−0.43**−0.50**−0.54**−0.62**(0.92)

Test of direct linkage between task and relationship conflicts

We conducted a series of nested model comparisons to test three competing hypotheses on the direct effects between task and relationship conflicts: (1) the task conflict → relationship conflict (Hypothesis 1a), (2) the relationship conflict → task conflict (Hypothesis 1b), and (3) the reciprocal relations between the two conflict types (Hypothesis 1c). The top section of Table 2 presents the results of the nested model tests and fit statistics for each model. The fit statistics indicated an adequate fit to the data for all three models. Next, we directly compared them by conducting nested model tests for evaluating the hypotheses.

Table 2. Fit statistics and chi-square difference tests for the direct linkage and the mediation models
Modelχ2dfComparisonΔχ2ΔdfCFIGFISRMR
  1. Note: N = 74 groups.

  2. *p < 0.01.

Direct linkage
 A. Autoregressive effects only10.33*2   0.950.940.08
 B. Task conflict → relationship conflict (Hypothesis 1a)8.42*1B vs. A1.9110.960.950.07
 C. Relationship conflict → task conflict (Hypothesis 1b)0.141C vs. A10.19*11.001.000.01
 D. Reciprocal effect (Hypothesis 1c)0.000D vs. C0.1411.001.000.00
Mediation effects
 E. No mediation effect32.36*2   0.860.880.08
 F. Task conflict → negative group affect → relationship conflict (Hypothesis 2a)12.00*1F vs. E20.36*10.950.940.07
 G. Relationship conflict → negative group affect → task conflict (Hypothesis 2b)0.591G vs. E31.77*11.001.000.01
 H. Saturated0.000H vs. G0.5911.001.000.00

We specified a base model (Model A) with auto-regressive but no cross-lagged paths as a basis for comparison with later models (Models B to D). In Models B through D, we added further paths to the base model: a cross-lagged structural path from task conflict T1 to relationship conflict T3 in Model B (Hypothesis 1a), a path from relationship conflict T1 to task conflict T3 in Model C (Hypothesis 1b), and reciprocal associations in Model D (Hypothesis 1c). The base model was compared with Models B and C, respectively, while Model C was compared with Model D. A significant difference in chi-squares indicates that the “larger” model with more parameters provides a better fit. On the other hand, a non-significant chi-square indicates that both models equally fit statistically. In this case, a “smaller” model with fewer parameter estimates is better because it is parsimonious (Mueller, 1996).

The addition of the task conflict T1 → relationship conflict T3 path to the base model (Model B vs. Model A) did not result in a significant increment in model fit, Δχ2 (1) = 1.91, ns. This result suggests no statistical evidence that task conflict T1 influences relationship conflict T3. Therefore, Hypothesis 1a was rejected. On the other hand, the addition of the relationship conflict T1 → task conflict T3 path to the base model (Model C vs. Model A) significantly improved model fit, Δχ2 (1) = 10.19, p < 0.01, supporting Hypothesis 1b. Finally, we compared the reciprocal effects model (Model D) with Model C. This comparison yielded no significant difference in chi-squares between the two model, Δχ2 (1) = 0.14, ns. This result suggests that Model C with fewer parameter estimates is more parsimonious than Model D. To further contrast these two models and evaluate the plausibility of each structural model, we examined the sign and significance of cross-lagged path coefficients in Figure 2. Although the chi-square different test suggested that Model C is the best fitting model, we presented Model D in order to more closely compare both cross-lagged paths between task and relationship conflict. The sign and significance of the parameters in Model D remained the same as in Model C, except for the task conflict T1 → relationship conflict T3 link that was not included in Model C.

thumbnail image

Figure 2. Standardized maximum likelihood parameter estimates for cross-lagged effects model. The dotted lines indicate statistically non-significant paths. (Note: Parameter estimates for the measurement model are not included, *p < 0.05, **p < 0.01.)

Download figure to PowerPoint

An examination of Figure 2 reveals that the cross-lagged path from relationship conflict T1 to task conflict T3 was significant (β = 0.43, p < 0.01), while the path from task conflict T1 to relationship conflict T3 was not significant (β = −0.05, ns). These findings are consistent with Hypothesis 1b, but suggest no statistical evidence for reciprocal associations between task and relationship conflict. Therefore, Hypothesis 1c was rejected.

Test of mediation effects

We conducted nested model comparisons to test two competing hypotheses on mediation effects of negative group affect on the relationships between task and relationship conflict: (1) the task conflict T1 → negative group affect T2 → relationship conflict T3 (Hypothesis 2a) and (2) the relationship conflict T1 → negative group affect T2 → task conflict T3 (Hypothesis 2b).

We first specified a baseline model (Model E) that included two autoregressive effects, two direct paths from each conflict type at T1 to the other type at T3, and two paths from negative group affect T2 to each conflict type at T3. Model E proposes no mediation effect in which both conflict types at T1 → negative group affect T2 are not posited. In Models F through H, we added a path from each conflict type at T1 to negative group affect T2 to the baseline model and then formed mediation effect models with both direct and indirect effects of independent variables: a path from task conflict T1 to negative group affect T2 in Model F (Hypothesis 2a), a path from relationship conflict T1 to negative group affect T2 in Model G (Hypotheses 2b), and both in Model H (the saturated model). The bottom section of the Table 2 presents the results of the nested model test and fit statistics for each model.

The fit statistics indicated that all mediation models fit the data sufficiently, and significant differences between the chi-squares of the baseline model (Model E) and the mediation models (Models F and G) supported superior fit for the mediation models, Δχ2 (1) = 20.36, p < 0.01 (Model F vs. E), Δχ2 (1) = 31.77, p< 0.01 (Model G vs. E). Next, we compared the two mediation models with the saturated model. The comparisons reveal that the saturated model fits the data better than Model F, which specified both direct and indirect effects of task conflict T1, Δχ2 (1) = 12.00, p < 0.01 (Model H vs. F), rejecting Hypothesis 2a. However, there was no significant difference between the saturated model and Model G, which specified both direct and indirect effects of relationship conflict T1, Δχ2 (1) = 0.59, ns (Model H vs. G). This result indicates that Model G with fewer parameters is more parsimonious than the saturated model, supporting Hypothesis 2b. We further examined the sign and significance of path coefficients in Figure 3. Likewise, although Model G is the best fitting model, we presented the saturated model.

thumbnail image

Figure 3. Standardized maximum likelihood parameter estimates for mediation effects model. The dotted lines indicate statistically non-significant paths. (Note: Parameter estimates for the measurement model are not included, *p < 0.05, **p < 0.01.)

Download figure to PowerPoint

An examination of Figure 3 provides support for an indirect effect relationship conflict T1 on task conflict T3, that is, relationship conflict T1 led to negative group affect T2 (β = 0.62, p < 0.01), which in turn evoked task conflict T3 (β = 0.41, p < 0.05). On the other hand, the significant direct effect of relationship conflict T1 on task conflict T3 (in Figure 2) became insignificant when controling for negative group affect T2 (reduced from β = 0.43, p < 0.01 to β = 0.18, ns).

Test of moderating effects

Hypotheses 3a and 3b, which predicted intragroup trust would moderate the effects of task conflict on relationship conflict and vice versa, were tested by applying a regression technique. We used centered predictors to facilitate testing interaction effects (Aiken & West, 1991).

Based on previous studies on task and relationship conflict, we included group size, gender diversity, race diversity, and perceived value diversity T3 to control for possible confounding effects. Table 3 presents the results of testing Hypotheses 3a and 3b. Hypothesis 3a was supported as the inclusion of the interaction term accounted for an additional four percent of the variance in relationship conflict T3, ΔR2 = 0.04, p < 0.05. The interaction term (task conflict T1 × intragroup trust T1) was significant and in the predicted direction (b = −0.55, p < 0.05). For a more conservative test, we controlled the testretest reliability of the dependent measure (i.e., relationship conflict T1). The interaction effect was reduced but still marginally predictive for relationship conflict T3, b = −0.40, p = 0.09. However, Hypothesis 3b was not supported because the inclusion of the relationship conflict T1 × intragroup trust T1 interaction did not account for a significant incremental variance in task conflict T3R2 = 0.02, ns).

Table 3. Results of regression analyses for the moderating effect of intragroup trust
VariablesRelationship conflict T3Task conflict T3
Step 1Step 2Step 1Step 2
  • Note: Unstandardized regression coefficients are presented. N = 74 groups; T1, Time 1; T2, Time 2; T3, Time 3.

  • *

    p < 0.05;

  • **

    p < 0.01.

Group size0.100.090.030.03
Gender diversity0.440.450.080.13
Race diversity0.030.10−0.10−0.13
Value diversity T30.50**0.52**0.32**0.31**
Task conflict T10.230.25  
Relationship conflict T1  0.50**0.55**
Trust T1−0.23*−0.27*−0.03−0.01
Task conflict T1 × trust T1 −0.55*  
Relationship conflict T1 × trust T1   0.34
R2 (Adjusted R2)0.52 (.48)0.56 (.51)0.47 (.43)0.49 (.44)
F12.18**11.82**10.03**9.14**
ΔR2 0.04* 0.02

To probe the nature of the significant interaction, we plotted the slopes of the simple regression of relationship conflict T3 on task conflict T1 according to the levels of trust in groups. Following Cohen and Cohen (1983), we defined the average-, low-, and high-trust groups, corresponding to the mean of intragroup trust T1 scores, one standard deviation below, and one standard deviation above the mean, respectively. Figure 4 shows that earlier task conflict is more strongly related to later relationship conflict at a low-trust group.

thumbnail image

Figure 4. Interaction effect of intragroup trust on the association between task conflict at time

Download figure to PowerPoint

Supplemental analyses

Although our results provided support for the moderating role of intragroup trust on the earlier task conflict-later relationship conflict relations, we did not include the mediation of negative group affect on the relations in the analysis. One may argue that intragroup trust may moderate the effects of earlier task conflict on subsequent native group affect. The presence of intragroup trust could prevent task conflict from triggering negative group affect and thereby attenuates relationship conflict. To investigate this issue more closely, we conducted a supplemental analysis where mediation and moderation effects were combined.1 The inclusion of both mediation and moderation effects in a single model would provide a better understating of the task-relationship conflict relations.

As shown in Figure 5, the key question in the analysis is whether or not intragroup trust moderated on (a) the task conflict T1 → negative group affect T2, or (b) the task conflict T1 → relationships conflict T3 linkages. Following Muller, Judd, and Yzerbyt's (2005) framework for combining moderation and mediation, we tested the three hierarchical linear regression equations. Table 4 specifies these equations. In Equation 1 which predicts relationship conflict T3, we entered the control variables of group size, gender diversity, race diversity, and value diversity T3 at Step 1, and the interaction term (task conflict T1 × intragroup trust T1) along with the centered predictors at Step 2. In Equation 2 which predicts negative group affect T2, the control variables of group size, gender diversity, race diversity, and value diversity T1 were entered at Step 1, and the centered task conflict T1, intragroup trust T1, and the interaction term (task conflict T1 × intragroup trust T1) were entered at Step 2. Equation 3 is identical to Equation 1, except that negative group affect T2 is now added at Step 2.

thumbnail image

Figure 5. Models combining the moderation and the mediation. Panel (a) illustrates the moderating effect of intragroup trust at Time 1 on the task conflict at Time 1 → negative group affect at Time 2. While panel (b) illustrates the moderating effect of intragroup trust at Time 1 on the task conflict at Time 1 → relationship conflict at Time 3

Download figure to PowerPoint

Table 4. Results of regression analyses for the mediation and moderation model
VariablesRelationship conflict T3 (Equation 1)Negative group affect T2 (Equation 2)Relationship conflict T3 (Equation 3)
  • Note: Unstandardized regression coefficients are presented. N = 74 groups; T1, Time 1; T2, Time 2; T3, Time 3.

  • p < 0.10;

  • *

    p < 0.05;

  • **

    p < 0.01.

Controls
Group size0.130.090.15*0.070.130.08
Gender diversity0.340.45−0.38−0.190.340.48
Race diversity0.060.10−0.020.100.060.06
Value diversity T1  0.74**0.15  
Value diversity T30.71**0.52**  0.71**0.37**
Task conflict T1 0.25 0.31* 0.20
Trust T1 −0.27** −0.39** −0.18
Task conflict T1 × trust T1 −0.55* 0.02 −0.55*
Negative group affect T2     0.31*
R2 (Adjusted R2)0.46 (.43)0.56 (.51)0.34 (.30)0.46 (.41)0.46 (.43)0.59 (.54)
F14.68**11.82**8.86**8.17**14.68**11.74**
ΔR2 0.10** 0.13** 0.13**

Table 4 reveals that the interaction term (task conflict T1 × intragroup trust T1) was significant while controling for negative group affect T2 (b = −0.55, p < 0.05) in Equation 3. However, the interaction term was not predictive for negative group affect T2 in Equation 2 (b = 0.02, ns). Thus, we concluded that intragroup trust moderated the direct effect of earlier task conflict on later relationship conflict, but not on the task conflict-negative group affect relation. This finding confirms Hypotheses 3a and is consistent with the findings of Simons and Peterson (2000).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

The primary purpose of this study was to investigate the patterns of causal linkage between task conflict and relationship conflict. We comprehensively derived possible conceptual models that specify the relationships between the two types of conflict and drew several research hypotheses on causality. To investigate the causal associations thoroughly, we employed a longitudinal panel design in which both types of conflicts were measured at two points in time and then performed a series of tests.

Results of testing direct causality demonstrated that earlier relationship conflict was significantly linked to later task conflict, but the opposite direction was not. Further investigation on the mediation models showed that relationship conflict triggered task conflict through negative group affect.

Testing the moderating effect of intragroup trust produced a consistent result with previous studies: The effect of task conflict on relationship conflict was significant for groups with a lower level of trust (Peterson & Behfar, 2003; Simons & Peterson, 2000). Earlier task conflict was more likely to cause a subsequent relationship conflict when group members did not trust one another strongly. On the other hand, intragroup trust was not found to moderate the reverse direction.

In sum, our findings suggest that it is somewhat unavoidable that relationship conflict bleeds over into task conflict in organizational settings; Personal animosity naturally triggers disagreements in task-related issues. On the other hand, task conflict may not be transferred into relationship conflict when intragroup trust is high. Furthermore, as seen in Figure 2, the effects of early relationship conflict on later relationship conflict (β = 0.75) and task conflict (β = 0.43) are much stronger than those of early task conflict. Given these findings, therefore, we may argue that relationship conflict should be managed well during the early stage of a group's development; otherwise, its influence to subsequent conflicts can be detrimental. With respect to task conflict, its impact on group outcome depends on group context and group process variables: When it is supported with such process as constructive controversy built on intragroup trust, it is not easily transformed into relationship conflict.

Our findings may provide a plausible explanation for why the majority of existing empirical research presents the negative consequences of task conflict. That is, a considerable amount of task conflict can erupt as a consequence of negative group affect resulting from relationship conflict. In this sense, we concur with Schwenk's (1990) notion that affect-free cognitive conflict is scarce in reality. Therefore, we argue that task conflict may be detrimental to team functioning not because of its destructive tendencies but because it is often the result of relationship conflict; interpersonal animosity is expressed as task disagreements. In this study, however, we were not able to test this argument because of lack of performance data. Before making a hasty conclusion on the ramifications on performance of task conflict in an organization, we need to investigate thoroughly when the two types of conflict are correlated and how we can separate the two. Future studies need to focus more on investigating the mechanisms and conditions that can reduce the dysfunctional effects of relationship conflict on task conflict.

Contributions

Although it is well-recognized that task conflict and relationship conflict are related, how they are related is relatively studied less. Our study provides insight into the pattern of associations between the two variables by showing evidence on the direction of causality between them. Our findings provide important insights to conflict literature by indicating that the direction of causality is reciprocal for some specific groups, particularly those with lower trust among members. However, for groups with higher levels of trust, relationship conflict can cause subsequent task conflict, whereas task conflict is not transformed into subsequent relationship conflict.

The result of testing cross-lagged effects reveals that relationship conflict is apt to cause task conflict, which suggests a promising research direction for future studies. Previous studies have exclusively tested the task conflict → relationship conflict linkage, but not the opposite. This was mainly because researchers were less interested in the relationship conflict → task conflict path or cross-sectional data were inadequate to test both directions. Our findings suggest that the relationship conflict → task conflict path is a more universalistic phenomenon and merits thorough investigation. As this study presents, the relationship conflict → task conflict causal flow has a valid theoretical background and is empirically significant.

We would like to mention several advantages in considering relationship conflict as a precedent of task conflict. First, relationship conflict is known to be related to individual differences such as personality (Asendorpf & Wilpers, 1998; Varela et al., 2008), attitude, or visible demographic characteristics (De Dreu & Van Vianen, 2001; Pearsall, Ellis, & Evans, 2008). In contrast, less is known about the source of task conflict (e.g., where it comes from, how it is proscribed or promoted, or who are more likely to perceive it). Therefore, if we start conflict management interventions from the relationship conflict, we are able to draw more implications that will help prevent or manage an appropriate level of conflict through direct interventions on the source of relationship conflict, from which indirect interventions on task conflict are also possible.

The result of moderation analysis provides important implications for trust literature. Our findings provide further support to an argument proposed by Dirks and Ferrin (2001) that, in many situations, trust should be expected to have a moderated rather than a direct effect on desired organizational outcomes. Second, although trust has been linked to a range of desired organizational outcomes (e.g., Dirks & Ferrin, 2001; Kramer, 1999), we are only beginning to understand its effects on task and relationship conflict. Our results indicate that trust can be expected to influence relationship conflict (a moderated effect), but not task conflict.

Finally, for practitioners to make use of research findings on the linkage between task conflict and relationship conflict, it is critical to know which is the determinant or the consequence of a particular type of conflict. Groups suffering from a negative spiral of task and relationship conflicts are likely to be those who are low in trust. Our results indicate that if trust is repaired in these groups, then the vicious cycle between task and relationship conflict can be broken due to the trustworthiness-cooperation spiral (Ferrin, Bligh, & Kohles, 2008). In addition, managers should not expect to reduce relationship conflict by focusing on improving the conflicts revolving on tasks because our results suggest that task conflict is unlikely to influence relationship conflict. Instead, it may be necessary to address interpersonal-related conflict among group members directly. Unless relationship conflict is handled well, its detrimental effect is more likely to be transformed into task conflict.

Limitations and future research directions

Although the present study attempted to shed light on the existing body of conflict literature by suggesting a rigorous prediction of the causal link from relationship conflict to task conflict, this study is not without limitations.

First, although a longitudinal panel design is often considered the preferred method to explore causal relations among variables, it is still limited in eliminating alternative interpretations attributable to a third or unmeasured variable. For example, social desirability or common method variance remains an unknown factor that could potentially influence causal relations between study variables. We tried to minimize the effect of unspecified variables by including age, sex, tenure, and value diversity as control variables and by allowing residual correlations in the test models. Replication of this study through a research design that enables maximal control over other factors (e.g., field experiment) is indeed desirable for future studies.

Second, self-report measures have been used exclusively in previous research on task and relationship conflicts. We also employed Jehn's (1994, 1995) self-report measures for this study. As Torrance (1957) noted, however, individuals tend to perceive all disagreements as evidence of personal rejections whether they are personal relationship- or task-related (318). Despite empirical separation of the two factors, this general tendency of human perception (i.e., difficulty in differentiating relationship conflict from task conflict) may have caused inflated correlations between the two constructs and could be an explanation for the high intercorrelations in existing research that relies on self-report measures. Thus, future researchers are encouraged to employ methods other than self-report measures.

Third, among many conditions that possibly influence the transformation between relationship conflict and task conflict, we tested only one possible moderator (i.e., trust). Future study merits identifying other moderating variables that may amplify or suppress the relationship. One promising research area is the role of conflict-relevant interactional norms (Yang & Mossholder, 2004). It is expected that some behavioral norms may reduce the strength or speed of relationship conflict transformed into task conflict. For example, under the influence of group norm that is intolerant to conflict among members or that proscribes open debates, group members are more likely to avoid situations that might bring conflict, whether it is a personal bicker or a task-related argument. If any type of conflict happens, then even appropriate interactions among group members will be halted, which blocks one type of conflict mutating into the other. In addition to its effect on the relationship conflict → task conflict flow, conflict-relevant norms may directly influence the magnitude of members' perception of conflict. For example, a group norm that promotes constructive controversy (Tjosvold, 1985) is more likely to encourage group members to accept others' perspectives without feelings of being threatened, which leads to weaker conflict perception. Thus far, there has been little empirical research on the role of group norms on conflict-relevant interactions. Further studies on the role of conflict-relevant interactional norms or conflict management strategies (Behfar, Peterson, Mannix, & Trochim, 2008; DeChurch, Hamilton, & Haas, 2007) will extend our understanding of the nature of the association between task and relationship conflict.

Fourth, the nature of the sample and the tasks the student groups performed may limit the generalization of our findings. Although the student groups worked like “real” teams while engaged in group projects for 13 weeks, they may have different characteristics compared with actual work teams in organizational settings. The student groups in this study were relatively homogeneous in terms of education, age, work experience, power status, and so on. They were also newly formed groups rather than already existing groups where the group members have a history of working together. The setting of the present study provides an empirical baseline for exploring how the relationship between task conflict and relationship conflict changes over time in initially formed groups because a newly formed group is relatively free from established relationships among its members. However, one may argue that causal directions may depend on the nature of their past work experiences together, which was not considered in this study. In particular, member familiarity affects task-relationship conflict associations. Familiar group members are more likely than strangers to form favorable first impressions, trust one another (Davis & Todd, 1985), accept other members' ideas at face value (Uzzi, 1996), and be comfortable disagreeing with one another (Gruenfeld, Mannix, Williams, & Neale, 1996). Furthermore, the groups in our study differ from existing work groups in the sense that they had a clearly defined life. If relationships among group members persist over time without delineated deadlines, the escalation of conflict is likely to happen, especially in low-performing teams (e.g., Jehn & Mannix, 2001) because one conflict type (either task or relationship conflict) is apt to stimulate the other conflict type. Thus, our findings are likely to generalize to a limited set of work teams in organizations (e.g., ad-hoc project teams) not necessarily concerned with establishing social relations and anticipating future relationships among team members.

Fifth, past performance feedback which was not considered in our study is likely to trigger emotional responses among group members and affect subsequent group interactions. Peterson and Behfar (2003) found that poor performance feedback increased later task and relationship conflict in groups. One may argue that positive feedback intervention may prevent task conflict from evolving into relationship conflict and vice versa. Therefore, future studies need to replicate our findings in teams where members have previously worked together and received performance feedback.

Despite the limitations, our study makes a contribution to the literature on task and relationship conflict by suggesting a causal flow from relationship conflict to task conflict. We hope this study will help advance our understanding of the dynamic nature of conflict in groups.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

For valuable feedback on previous drafts, the authors thank Dean Tjosvold, and three anonymous reviewers. Authors also thank the late James Meindl, Donald Ferrin, Tony Simons, Corinne Coen, and Fred Dansereau for valuable feedback on earlier drafts of this work. This work was supported by National Research Foundation of Korea Grant funded by the Korean Government (332-2006-1-B00118).

  • 1

    We thank anonymous reviewers for suggesting this analysis.

Biographical Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

Kyoosang Choi is an Assistant Professor at Sookmyung Women's University, Seoul, South Korea. He received his PhD in Organization and Human Resources from the State University of New York at Buffalo. His research interests include conflict management, group cognition, and leadership.

Biographical Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References

Bongsoon Cho is an Associate Professor at Sogang University, Seoul, South Korea. He received his PhD in Organization and Human Resources from the State University of New York at Buffalo. His research explores the role of social identity in organizational context such as mergers and acquisitions, expatriate assignment, intergroup conflict, and diversity management.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Models of Task and Relationship Conflict
  5. Methods
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. References
  • Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.
  • Amason, A. C. (1996). Distinguishing the effects of functional and dysfunctional conflict on strategic decision making: Resolving a paradox for top management groups. Academy of Management Journal, 39, 123148.
  • 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, 495516.
  • Anderson, S. E., & Williams, L. J. (1992). Assumptions about unmeasured variables with studies of reciprocal relationships: The case of employee attitudes. Journal of Applied Psychology, 77, 638650.
  • Arbuckle, J. L. (2007). AMOS 16.0 User's guide. Chicago: SmallWaters.
  • Armstrong, J. S., Brodie, R. J., & Parsons, A. G. (2001). Hypotheses in marketing science: Literature review and publication audit. Marketing Letters, 12, 171187.
  • Asendorpf, J. B., & Wilpers, S. (1998). Personality effects on social relationships. Journal of Personality and Social Psychology, 74, 15311544.
  • Barsade, S. G. (2002). The ripple effect: Emotional contagion and its influence on group behavior. Administrative Science Quarterly, 47, 644675.
  • Bartel, C. A., & Saavedra, R. (2000). The collective construction of work group moods. Administrative Science Quarterly, 45, 197231.
  • Behfar, K., Peterson, R., Mannix, E. A., & Trochim, W. (2008). The critical role of conflict resolution in teams: A close look at the links between conflict type, conflict management strategies, and team outcomes. Journal of Applied Psychology, 93, 170180.
  • Berscheid, E. (1983). Emotion. In H. H. Kelley, E. Berscheid, A. Christensen, J. H. Harvey, T. L. Huston, G. Levinger, E. McClintock, L. A. Peplau, & D. R. Peterson (Eds.), Close relationships (pp. 110168). New York: W.H. Freeman and Company.
  • Blau, P. M. (1977). Inequality and heterogeneity. New York: The Free Press.
  • Carnevale, P. J. D., & Isen, A. M. (1986). The influence of positive affect and visual access on the discovery of integrative solutions in bilateral negotiation. Organizational Behavior and Human Decision Processes, 37, 113.
  • Chatman, J. A., Polzer, J. T., Barsade, S. G., & Neale, M. A. (1998). Being different yet feeling similar: The influence of demographic composition and organizational culture on work process and outcomes. Administrative Science Quarterly, 43, 749780.
  • Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analyses for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
  • Cooper, J., & Fazio, R. H. (1979). The formation and persistence of attitudes that support intergroup conflict. In W. G. Austin, & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 149159). Monterey, CA: Brooks/Cole.
  • Davis, K. E., & Todd, M. J. (1985). Assessing friendships: Prototypes, paradigm cases, and relationship description. In S. Duck, & D. Perlman (Eds.), Understanding personal relationships: Sage series in personal relationships (Vol. 1; pp. 1737). Beverly Hills: Sage.
  • De Dreu, C. K. W. (2008). The virtue and vice of workplace conflict: Food for (pessimistic) thought. Journal of Organizational Behavior, 29, 518.
  • De Dreu, C. K. W., & Van Vianen, A. E. M. (2001). Managing relationship conflict and the effectiveness of organizational teams. Journal of Organizational Behavior, 22, 309328.
  • De Dreu, C. K. W., & Weingart, L. R. (2003). Task versus relationship conflict, team performance, and team member satisfaction: A meta-analysis. Journal of Applied Psychology, 88, 741749.
  • DeChurch, L. A., Hamilton, K. L., & Haas, C. (2007). Effects of conflict management strategies on perceptions of intragroup conflict. Group Dynamics: Theory, Research, and Practice, 11, 6678.
  • Dirks, K. T., & Ferrin, D. L. (2001). The role of trust in organizational settings. Organization Science, 12, 450467.
  • Eisenhardt, K. M., & Bourgeois, L. J. (1988). Politics of strategic decision making in high-velocity environments. Academy of Management Journal, 32, 543576.
  • Ferrin, D. L., Bligh, M. C., & Kohles, J. C. (2008). It takes two to tango: An interdependence analysis of the spiraling of perceived trustworthiness and cooperation in interpersonal and intergroup relationships. Organizational Behavior and Human Decision Processes, 107, 161178.
  • Gamero, N., Gonzalez-Roma, V., & Peiro, J. M. (2008). The influence of intra-team conflict on work teams' affective climate: A longitudinal study. Journal of Occupational and Organizational Psychology, 81, 4769.
  • George, J. M. (1990). Personality, affect, and behavior in groups. Journal of Applied Psychology, 75, 107116.
  • George, J. M., & Brief, A. P. (1992). Feeling good-doing good: A conceptual analysis of the mood at work-organizational spontaneity relationship. Psychological Bulletin, 112, 310329.
  • Georgopolous, B. S. (1986). Organizational structure, problem solving, and effectiveness. San Francisco: Jossey-Bass.
  • Gero, A. (1985). Conflict avoidance in consensual decision processes. Small Group Behavior, 16, 487499.
  • Gersick, C. J. G. (1988). Time and transition in work teams: Toward a new model of group development. Academy of Management Journal, 31, 941.
  • Goldman, M., Cowles, O., & Florez, C. (1983). The halo effect of an initial impression upon speaker and audience. Journal of Social Psychology, 120, 197201.
  • Gouaux, C. (1971). Induced affective states and interpersonal attraction. Journal of Personality and Social Psychology, 20, 3743.
  • Gruenfeld, D. H., Mannix, E. A., Williams, K. Y., & Neale, M. A. (1996). Group composition and decision making: How member familiarity and information distribution affect process and performance. Organizational Behavior and Human Decision Processes, 67, 115.
  • Guetzkow, H., & Gyr, J. (1954). An analysis of conflict in decision-making groups. Human Relations, 7, 367381.
  • Hobman, E. V., & Bordia, P. (2006). The role of team identification in the dissimilarity–conflict relationship. Group Processes & Intragroup Relations, 9, 483507.
  • Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 155.
  • Isen, A. M. (1970). Success, failure, attention, and reaction to others: The warm glow of success. Journal of Personality and Social Psychology, 15, 294301.
  • Isen, A. M., & Levin, P. F. (1972). The effect of feeling good on helping: Cookies and kindness. Journal of Personality and Social Psychology, 21, 384388.
  • James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group reliability with and without response bias. Journal of Applied Psychology, 69, 8598.
  • Janis, I. L. (1972). Victims of groupthink: Psychological studies of foreign policy decisions and fiascoes. Boston: Houghton Mifflin.
  • Jehn, K. (1994). Enhancing effectiveness: An investigation of advantages and disadvantages of value-based intragroup conflict. International Journal of Conflict Management, 5, 223238.
  • Jehn, K. (1995). A multimethod examination of the benefits and detriments of intragroup conflict. Administrative Science Quarterly, 40, 256282.
  • Jehn, K. (1997). A qualitative analysis of conflict types and dimensions in organizational groups. Administrative Science Quarterly, 42, 530557.
  • Jehn, K., & Bendersky, C. (2003). Intragroup conflict in organizations: A contingency perspective on the conflict-outcome relationship. In R. M. Kramer, & B. M. Staw (Eds.), Research in Organizational Behavior (Vol. 25; pp. 187242). Greenwich, CT: JAI Press.
  • Jehn, K., & Mannix, E. (2001). The dynamic nature of conflict: A longitudinal study of intragroup conflict and group performance. Academy of Management Journal, 44, 238251.
  • Jehn, K., Northcraft, G. B., & Neale, M. A. (1999). Why differences make a difference: A field study of diversity, conflict, and performance in workgroups. Administrative Science Quarterly, 44, 741763.
  • Jones, T. S. (2000). Emotional communication in conflict: Essence and impact. In W. Eadie, & P. Nelson (Eds.), The language of conflict and resolution (pp. 81104). Thousand Oaks, CA: Sage.
  • Kearney, E., Gebert, D., & Voelpel, S. C. (2009). When and how diversity benefits teams: The importance of team members' need for cognition. Academy of Management Journal, 52, 581598.
  • Kelley, H. H. (1950). The warm-cold variable in first impressions of persons. Journal of Personality, 18, 431439.
  • Klein, K. J., & Kozlowski, S. W. J. (2000). From micro to meso: Critical steps in conceptualizing and conducting multilevel research. Organizational Research Methods, 3, 211236.
  • Kramer, R. M. (1999). Trust and distrust in organization: Emerging perspectives, enduring questions. Annual Review Psychology, 50, 569598.
  • Lindner, E. G. (2006). Emotion and conflict: Why it is important to understand how emotions affect conflict and how conflict affects emotions. In M. Deutsch, P. T. Coleman, & E. C. Marcus, second (Eds.), The handbook of conflict resolution: Theory and practice. (pp. 268293) San Francisco: Jossey-Bass.
  • Milberg, S., & Clark, M. S. (1988). Moods and compliance. British Journal of Social Psychology, 27, 7990.
  • Mooney, A. C., Holahan, P. J., & Amason, A. C. (2007). Don't take it personally: Exploring cognitive conflict as a mediator of affective conflict. Journal of Management Studies, 44, 733758.
  • Mueller, R. O. (1996). Basic principles of structural equation modeling. New York: Springer.
  • Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When moderation is mediated and mediation is moderated. Journal of Personality and Social Psychology, 89, 852863.
  • Nester, J., Kutner, M. H., Nachtscheim, C. J., & Wasserman, W. (1996). Applied linear statistical models (4th ed.) Boston: McGraw-Hill.
  • Pearsall, M. J., Ellis, A. P., & Evans, J. M. (2008). Unlocking the effects of gender faultlines on team creativity: Is activation the key? Journal of Applied Psychology, 93, 225234.
  • Pelled, L. H., Eisenhardt, K. M., & Xin, K. R. (1999). Exploring the black box: An analysis of work group diversity, conflict, and performance. Administrative Science Quarterly, 44, 128.
  • Peterson, R. S., & 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, 102112.
  • Pirola-Merlo, A., Hartel, C., Mann, L., & Hirst, G. (2002). How leaders influence the impact of affective events on team climate and performance in R&D teams. The Leadership Quarterly, 13, 561581.
  • Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879903.
  • Priem, R. L., & Price, K. H. (1991). Process and outcome expectations for the dialectical inquiry, devil's advocacy, and consensus techniques of strategic decision making. Group and Organization Studies, 16, 206225.
  • Pruitt, D. G., & Rubin, J. Z. (1986). Social conflict: Escalation, stalemate, and settlement. New York: Random House.
  • Ross, R. S. (1989). Small groups in organizational settings. Englewood Cliffs, NJ: Prentice-Hall.
  • Schwenk, C. R. (1990). Conflict in organizational decision making: An exploratory study of its effects in for-profit and not-for-profit organizations. Management Science, 36, 436448.
  • Simons, T., & Peterson, R. (2000). Task conflict and relationship conflict in top management teams: The pivotal role of intragroup trust. Journal of Applied Psychology, 85, 102111.
  • Simons, T., Pelled, L. H., & Smith, K. A. (1999). Making use of difference: Diversity, debate, and decision comprehensiveness in top management teams. Academy of Management Journal, 42, 662673.
  • Spector, P. E. (2008). Introduction: Conflict in organizations. Journal of Organizational Behavior, 29, 34.
  • Suls, J., Martin, R., & David, J. P. (1998). Person-environment fit and its limits: Agreeableness, neuroticism, and emotional reactivity to interpersonal conflict. Personality and Social Psychological Bulletin, 24, 8898.
  • Thorndike, E. L. A. (1920). A constant error in psychological ratings. Journal of Applied Psychology, 4, 2529.
  • Tidd, S. T., McIntyre, H. H., & Friedman, R. A. (2004). The importance of role ambiguity and trust in conflict perception: Unpacking the task to relationship conflict linkage. International Journal of Conflict Management, 15, 364380.
  • Tjosvold, D. (1985). Implications of controversy research for management. Journal of Management, 11, 221238.
  • Tjosvold, D. (1992). Team organization: An enduring competitive advantage. New York: Wiley.
  • Tjosvold, D. (2008). The conflict-positive organization: It depends upon us. Journal of Organizational Behavior, 29, 1928.
  • Torrance, E. P. (1957). Group decision making and disagreement. Social Forces, 35, 314318.
  • Totterdell, P. (2000). Catching moods and hitting runs: Mood linkage and subjective performance in professional sport teams. Journal of Applied Psychology, 85, 848859.
  • Uzzi, B. (1996). The Sources and consequences of embeddedness for the economic performance of organizations: The network effect. American Sociological Review, 61, 674698.
  • Van de Ven, A. H., & Ferry, D. (1980). Measuring and assessing organizations. New York: Wiley Interscience.
  • Varela, O. E., Burke, M. J., & Landis, R. S. (2008). A process model of emotional conflict: Emergence and dysfunctional effect in groups. Group Dynamics: Theory, Research, and Practice, 12, 112126.
  • Weiss, H. M., & Cropanzano, R. (1996). Affective events theory: A theoretical discussion of the structure, causes and consequences of affective experiences at work. In B. M. Staw, & L. L. Cummings (Eds.), Research in Organizational Behavior (Vol. 19; pp. 174). Greenwich, CT: JAI Press.
  • Welty, J. D. (1983). Debate vs. consensus in the boardroom. Directors & Boards, 8, 3437.
  • West, S. G. (1975). Personalism and cognitive labels as determinants of attitude attribution. Memory & Cognition, 3, 496500.
  • Williams, L. J., & Hazer, J. T. (1986). Antecedents and consequences of satisfaction and commitment in turnover models: A reanalysis using latent variable structure equation methods. Journal of Applied Psychology, 71, 219231.
  • Yang, J., & Mossholder, K. W. (2004). Decoupling task and relationship conflict: The role of intragroup emotional processing. Journal of Organizational Behavior, 25, 589605.