To Agree or Not To Agree: The Effects of Value

Download PDF
0 downloads 0 Views 65KB Size Report
Comment
whether or not this diversity is a help or a hindrance to group outcomes (Jackson, 1992; Pelled, 1996). In this paper, we investigate value congruence and ...
To Agree or Not To Agree: The Effects of Value Congruence, Individual Demographic Dissimilarity and Conflict on Workgroup Outcomes

Karen A. Jehn Clint Chadwick Sherry M.B. Thatcher 2000 Steinberg Hall - Dietrich Hall University of Pennsylvania Philadelphia, PA 19104-6370 tel. 215 898-0525 fax 215 898-0401 [email protected]

Running head: Value Congruence and Demographic Dissimilarity We would like to thank the Dispute Resolution Research Center at Northwestern University for financial assistance. We would also like to thank Thomas Armentrout for his involvement in the project, Cindy Young for her data entry and analysis efforts, and Kelly Gable for her assistance in manuscript

preparation. Abstract To Agree or Not To Agree: The Effects of Value Congruence, Individual Demographic Dissimilarity and Conflict on Workgroup Outcomes

In this quasi-experimental study, we investigate value congruence and demographic dissimilarity among group members as factors which influence various types of conflict within workgroups. We also examine whether it is beneficial for members to be different or alike, to agree or disagree, in order to foster work group productivity. Results indicate that visible individual demographic differences (i.e., sex, age) increase relationship conflict, while informational demographic differences (i.e., education) increase task-focused conflict. Value congruence of members decreased both relationship and task conflict, and the specific content of the values held by members influenced performance. Specifically, both detail and outcome group value orientations increased objective performance; outcome, decisiveness, and stability orientations increased perceptions of high performance; and, both decisiveness and supportiveness orientations increased the satisfaction level of group members while a team orientation decreased individual member satisfaction in this sample.

To Agree or Not To Agree: The Effects of Value Congruence, Individual Demographic Dissimilarity and Conflict on Workgroup Outcomes There has been a debate in organizational research regarding whether or not agreement or disagreement within groups is advantageous. Conflict researchers (Amason, 1996; Amason & Schweiger, 1994; Jehn, 1994, 1995, 1997) have recently found that while relationship conflicts based on personality clashes and interpersonal dislike are detrimental to group functioning, task conflicts based on disagreements regarding the specific task content are beneficial in many situations. Another prominent, though less researched, controversy surrounds the concept of group member diversity and whether or not this diversity is a help or a hindrance to group outcomes (Jackson, 1992; Pelled, 1996). In this paper, we investigate value congruence and demographic dissimilarity among group members as factors which influence various types of conflict within workgroups. More specifically, we examine which sets of values are likely to demonstrate the greatest functional and dysfunctional effects in groups, and how the visibility of individual demographic and informational background characteristics affect conflict and performance within the group. In general, we examine the individual characteristic inputs and the outcomes of workgroups to determine whether or not it is beneficial for members to be different or alike, and whether or not they should agree or disagree in order to foster work group productivity. Value Congruence and Intragroup Conflict Values are individuals' fundamental beliefs regarding the desirability of behavioral choices (Enz, 1988; Rokeach, 1973). They reflect, for example, preferred ways to perform individual and group tasks such as being innovative, detail oriented, or relationship-focused. In this study, we examine value 3

congruence, which is defined as the degree to which all members of the group agree on values about group processes and group work, and its effects on intragroup conflict. In this study, we examine two types of intragroup conflict - relationship conflict and task conflict - based on the past work of Jehn (1995). While other labels have been used to describe various types of conflict (e.g., emotional and task, Jehn, 1994; cognitive and affective, Amason, 1996; Amason & Schweiger, 1994), Jehn’s recent work (1997) has distinguished emotion as a separate component of conflict that can be present in either a task or relationship conflict situation. Therefore, to be clear about the content area of the conflict without confounding the dimension of emotionality or affect (see Jehn, 1997 for a more thorough discussion of this), we use the labels “relationship conflict” and “task conflict.” Relationship conflicts are disagreements and incompatibilities among group members about personal issues that are not task-related, such as social events, gossip, and world news. This type of conflict often includes personality differences, animosity, and annoyance between individuals. Task conflicts are disagreements among group members’ ideas and opinions about the task being performed, such as disagreement regarding an organization’s current strategic position or determining the correct data to include in a report. The similarity of values among group members has been shown to influence the amount of conflict within the group (Jehn, 1994; Pelled, 1996). Because values are a guide for behavioral choices, group members who share similar values are more likely to agree about group actions such as goals, tasks, and procedures, thus reducing task conflict. Moreover, because values can act as perceptual filters, members with similar values are more likely to prioritize and interpret group problems and events in similar ways, further reducing task conflict. Similarity of values also reduces relationship conflict by 4

increasing the degree to which group members identify with one another. We propose that when members have similar values regarding the workgroup and its goals, they will have similar opinions based on those common beliefs (Jehn, 1994), therefore reducing the amount of both intragroup taskand relationship-focused conflict. Our model is shown in Figure 1 and our first two hypotheses follow: H1:

Group value congruence will decrease relationship conflict in workgroups.

H2:

Group value congruence will decrease task conflict in workgroups. ------------------------------------------------------Insert Figure 1 about here -------------------------------------------------------

Diversity Within Groups: Individual Demographic Dissimilarity While much research exists on the effects that one type of diversity has upon the performance of groups (e.g., Zenger & Lawrence, 1989; Tsui & Egan, 1994), fewer studies have attempted to gauge the relative importance of the various attributes of a group’s demographic composition (for a comprehensive review of organizational demography studies see Tsui, Egan, and Xin, 1995) and the impact on individual level outcomes. In addition, many studies have attempted to show the effects of group diversity on performance without taking into account any intervening processes, such as conflict (e.g., Bantel & Jackson; 1989; Kent & McGrath, 1969; Ancona & Caldwell, 1992; Watson, Kumar, & Michaelsen, 1993). Pelled (1996) develops a persuasive argument that links demographic diversity to group performance as mediated by task and relationship-focused conflict, but there is no empirical evidence to support the model. While we do not test this model directly, it does provide the conceptual foundation for the relationships (discussed below) regarding demographic diversity and conflict in the 5

current study. For example, Pelled’s model focuses on groups and group outcomes; while we are primarily concerned with individuals and the effect that dissimilarity to other group members has on group interaction and outcomes. Demographic attributes usually consist of age, sex, race, education level, work experience, and organization tenure. Recent research in this area has classified these attributes on two dimensions: visibility and job-relatedness (Pelled, 1996; Tsui, Egan, & O’Reilly, 1992). In this paper, visible demographic characteristics refer to age, sex, and race; informational demographic characteristics refer to differences among individuals related to education level, work experience, and organization tenure. The distinction between these two demographic types is important because observable characteristics may not be as relevant to the given task but they shape people, perceptions and behaviors (Pelled, 1996; Dienesch & Liden, 1986). In addition, people use visible demographic characteristics for categorization processes, which can cause hostility, anxiety, and stereotyping (Tsui, Egan, & O’Reilly, 1992). Furthermore, the initial perceptions made by and about group members often signal the quality and type of the future relationships among the team members which tend to persist over time (Tsui, Xin, & Egan, 1995). These differences have an increasing impact on a group member as the number and degree of differences on visible demographic attributes within the group increases. Thus, the degree to which an individual differs from group members in visible demographic characteristics can have profound effects on the amount of conflict that the person experiences within the group. Previous studies have shown that relationship conflict is more likely to be motivated by visible demographic characteristics than informational demographic characteristics (Alagna, Reddy, & Collins, 1982). This occurs because categorizing people based on visible attributes draws attention to features 6

of an individual which may be irrelevant to the task at hand. For example, referring to someone’s age when discussing an issue may make some people uncomfortable and create resentment among group members. As suggested by the similarity attraction paradigm (Byrne, 1971), the more different an individual is from other group members on a given demographic characteristic, the more likely they are to perceive relationship conflict. This occurs not only because people are attracted to those who are similar to themselves, but because they assume that people like them share the same values and worldview that they do (McPherson & Smith-Lovin, 1987). They also assume that similar others are easier to work and communicate with as well as believing they are more trustworthy (Brewer, 1981; Kanter, 1977). Conversely, individuals tend to generate negative evaluations of dissimilar others. Thus, we would expect that the more the visible demographic characteristics of group members differ, the more they are likely to perceive relationship conflict. On the other hand, attributes such as educational level and work experience often dictate how one thinks about and undertakes tasks. Therefore, informational demographic characteristics are more likely to bring about task-related conflict than are visible demographic characteristics (Ancona, 1990). Again, the number and degree of differences among individuals on informational demographic attributes raise the level of task conflict that those people experiences We suggest the following hypotheses: H3: Individual dissimilarity regarding visible demographic characteristics increases individual perceptions of relationship conflict in a work group. H4: Individual dissimilarity regarding informational demographic characteristics increases individual perceptions of task conflict in a work group. Conflict and Performance 7

By examining values, individual demographic characteristics, and intragroup conflict, we predict effects on three group outcomes: objective performance, perceptions of performance, and members’ satisfaction with their groups. Objective performance is the productivity of the group which can be measured by objective criteria (e.g., counting widgets, profitability, customer return). Perceived performance is defined as group members’ perceptions of how well they think they are performing; satisfaction is the degree to which the members are happy working in the group. Previous research (Amason & Schweiger, 1994; Jehn, 1994; Jehn 1995) has suggested that personal attacks and interpersonal disagreements within groups cause dissatisfaction among members and decrease the amount of individual effort put into completing the group task adequately. Members do not feel comfortable working in a group in which members are attacking one another. When relationship conflict is perceived in a group, feelings of dissatisfaction with the group experience will prevail. In addition, effort is misplaced on squabbling, avoiding, or resolving the interpersonal issues rather than focusing on task completion. Thus, work groups generally experience increasing process loss and withdrawal of individual members from the group as relationship conflict increases. In turn, process loss and withdrawal of group members from cooperative behaviors decreases opportunities for coordination of group tasks and synergies among group members in achieving their individually-assigned group tasks. Therefore, we predict that relationship conflict is negatively related to objective group performance. Moreover, because relationship conflict is unpleasant and can depress objective performance, group members’ perceptions of their group performance should decline as their perceptions of relationship conflict within their group rise. Negative emotional reactions to relationship 8

conflict will negatively bias members’ perceptions of outcomes. We therefore predict that: H5: Member perceptions of relationship conflict decrease objective group performance. H6: Member perceptions of relationship conflict decrease member perceptions of performance. H7: Member perceptions of relationship conflict decrease member satisfaction. Following the logic of Amason (1996) and Jehn (1994; 1995) we also predict that task conflict will affect performance in groups working on nonroutine tasks such as decision-making and consulting projects. Critical debate among members and open discussion regarding task issues increases group performance because members are more likely to offer and evaluate various solutions, thus reaching optimal decisions and outcomes (Jehn and Shah, 1997). Workgroups benefit from the increased debate about ideas and viewpoints. However, group members may still be bothered by the often contentious tone and harsh criticism which accompanies much task conflict (Jehn, 1997). When members feel discomfort with the group process and dissatisfaction with the group experience, they are more likely to believe that they have performed at a lower level than groups who enjoy the experience. This causes groups with high levels of task conflict to perform well objectively, but have low levels of satisfaction and perceptions of performance. In addition, group members who perceive low levels of task conflict may mistakenly believe their group has performed well, despite the inadequate levels of discussion, debate, and critical evaluation which increase performance (Shah & Jehn, 1993). H8: Member perceptions of task conflict increase objective group performance. H9: Member perceptions of task conflict decrease perceptions of performance. H10: Member perceptions of task conflict decrease member satisfaction. Value Content Intensity and Performance 9

The content of specific workgroup values, as well as member agreement or congruence on these values, also impacts objective performance, but we propose it does so directly rather than through relationship and task conflict. Value content intensity describes the relative emphasis an individual or group puts on different value dimensions.1 Value content refers to specific dimensions or types of values, such as being innovative or being detailed-oriented, which characterize group members’ sets of values (Schein, 1985), whereas value congruence is the degree of value similarity across group members. Previous research on group-level values and consensus has recognized the importance of both value congruence and value content intensity (Amason, 1996), and the potential liabilities of congruence, such as complacency and groupthink, (Enz, 1988; Janis, 1982), though clear distinctions between the two concepts of value congruence and value content intensity, such as those drawn here, are rare. In this quasi-experimental study, we investigate nine different value content dimensions: innovativeness, stability, detail orientation, outcome orientation, aggressiveness, supportiveness, reward orientation, team orientation, and decisiveness (for a more thorough discussion of the development of this typology see O'Reilly, Chatman, & Caldwell, 1991 and Chatman & Jehn, 1994). The innovation 1

For example, if a group value regarding innovation is measured by aggregating individual members' scores on a scale for this value, intensity and congruence would be approximated by the group's mean and standard deviation of these scores, respectively. Following the logic of Jehn (1994), intensity can only be measured in groups with moderate to high levels of congruence. It is illogical to consider group intensity at extremely low levels of congruence; if members do not agree on the important aspects of work values, they cannot have a high level of group value intensity.

10

dimension includes the specific values of being innovative, willing to experiment, and risk-taking. The stability dimension entails being careful, being rule-oriented, and valuing security of employment. The detail orientation dimension comprises values of being analytical, paying attention to detail, and being precise. The outcome orientation dimension includes values of being demanding, achievement orientation, and being results-oriented. Chatman and Jehn’s (1994) aggressiveness dimension contains values of being quick to take advantage of opportunities, being aggressive, and being competitive. The supportiveness dimension includes values of sharing information freely, being supportive, and offering praise for good performance (O’Reilly, Chatman, & Caldwell, 1991). The reward dimension comprises values of opportunities for professional growth, high pay for good performance, and fitting in, while the team orientation dimension contains being team oriented, working in collaboration with others, and autonomy. Lastly, the decisiveness dimension includes values of predictability, being decisive, and low levels of conflict (again, for a more in depth discussion of these content dimensions see O’Reilly, Chatman, & Caldwell, 1991; Chatman & Jehn, 1994). Given the nature of the task performed by the groups in this study (a nonroutine, interdependent group problem-solving task with a concrete deadline; Jehn, 1994; Shah & Jehn, 1993), we hypothesize that the following dimensions have positive impacts on groups’ objective performance: innovativeness, detail orientation, outcome orientation, reward orientation, team orientation, and decisiveness. Innovativeness is critical because consulting projects are often problem-solving tasks and are designed to provide new and unique insights to the customer. Detail orientation is important because time pressures and the demands of other projects often push employees to cut corners in ways that can negatively impact the quality of their projects. Thus strong emphasis on attention to detail can help 11

mitigate the negative effects of time pressures and lead to increased performance. Outcome orientation, reward orientation, and decisiveness increase task performance when a specific product needs to be created within a specific timeframe. In these situations, it will be detrimental for groups not to make concrete decisions about the specific product they are being rewarded to produce. Clearly, decisiveness will increase the likelihood that these decisions are made; reward and outcome orientations will improve the likelihood that these decisions will maximize objective performance evaluations, which are often primary organizational rewards and were an important outcome for the groups in this study. Finally, team-based consulting projects involve group tasks that require coordination and cooperation; therefore, a team orientation will positively affect groups’ objective performance. We propose the following regarding value content intensity and objective performance: H11: High group value content intensity on dimensions of innovativeness, detail orientation, outcome orientation, reward orientation, team orientation, and decisiveness are positively related to objective group performance. The factors which influence individuals’ perceptions of their performance are based on their implicit theories about the emphases required to succeed. In our study, the nonroutine nature of the group task creates ambiguity about the specific behaviors required by the task to succeed. Because group members in our study have little specific information upon which to base their implicit theories of performance, we do not expect systematic relationships between all of the value content dimensions and perceived performance. Instead, value dimensions which reflect a general emphasis on working hard and steadily, such as stability and attention to detail, will be positively related to perceived performance because they reflect prevailing values for a project of this type. The uncertainty of our groups’ task 12

environment suggests that factors linked with perceptions of self-efficacy, efficiency, and overall effort, such as stability, attention to detail, outcome orientation, aggressiveness, and decisiveness, will be positively related to individuals’ perceptions of their performance (Bandura, 1982). Additionally, the limited time period for our groups’ projects suggests that they will view an outcome orientation and decisiveness as important parts of a successful project. H12: High group value content intensity on dimensions of stability, detail orientation, outcome orientation, aggressiveness, and decisiveness are positively related to members’ perceptions of performance. Mechanisms emphasizing the affective nature of group processes relate certain value dimensions to individual satisfaction. For example, a group which emphasizes supportiveness should generate more emotional support for individual members, and this value dimension should be positively correlated with satisfaction. Similarly, team oriented groups are more likely to behave synergistically and in supportive ways, and decisive groups often feel more efficient and less ambiguous to their members. Both of these mechanisms reduce conflict and create a comfortable interpersonal climate within a team; therefore, these values are predicted to be positively related to member satisfaction. By the same reasoning, aggressive teams are more prone to interpersonal friction and will be more likely to sense dissatisfaction. Groups with high levels of aggressiveness are more likely to emphasize competitiveness over cooperation. The dissonance created by juxtaposing widespread aggressiveness with group tasks which ostensibly demand cooperativeness can lead to reduced satisfaction among group members: H13a: High group value content intensity on dimensions of supportiveness, team orientation, and decisiveness are positively related to satisfaction. 13

H13b: High group value content intensity on the dimension of aggressiveness is negatively related to satisfaction. Methods Sample and Procedure This quasi-experimental field study utilized functional groups performing a comparable task. The sample included eighty-eight teams of five participants (team size was held constant). Members were primarily full-time employees and enrolled as part-time students at two business schools and fulltime MBA students at a third. The average age of the members was 27.5 and 57% were male. The groups worked together as consulting teams for various organizations over a fourteen-week period. They were responsible for identifying an organizational problem, assessing the situation, developing a solution related to the strategic plan of the organization, and presenting recommendations for implementation. The teams worked together an average of 8.4 hours a week analyzing company records, attending organizational meetings, distributing group tasks, discussing strategic alternatives, developing recommendations and preparing a final report and presentation. The final product of the teams was a ten-page report. The values of the team members were assessed before groups were randomly formed, and the groups placed into low value consensus or high value consensus teams using a binary median split (Jehn, 1994). A questionnaire was administered to the team members at the completion of the project before they were given formal feedback on their project. Measures Individual demographic dissimilarity. We measured dissimilarity by creating a relational demography score (Tsui, Egan, & O’Reilly, 1992, p. 562): 14

“the square root of the summed squared differences between an individual SI’s value on a specific demographic variable and the value on the same variable for every other individual Sj in the sample for the work unit, divided by the total number of respondents in the unit (n).” For a more complete discussion of the formula and its origins, see Tsui, Egan, and O’Reilly (1992). The strength of this formula is that the relative amount of diversity that exists within a group can be determined for each variable. Therefore, for each individual in a group we derive four different dissimilarity variables (age, sex, nationality, and education) based on the relational demography scores. Unfortunately, we were not allowed to collect information on race or ethnicity. Value congruence and value content intensity. We measured value congruence using the Organizational Culture Profile (OCP) developed by O’Reilly, Chatman, and Caldwell (1991). The OCP is a Q-sort useful for identifying the central values of individuals and for comparing the similarity of values across group members. It consists of 54 items which participants sorted into 9 categories (“most important” to “most unimportant”) to reflect their individual values. To determine value congruence, the coefficient alpha was calculated based on the group members’ individual sorts of the 54 items (Jehn, 1994). The coefficient alpha refers to the degree to which group members have similar values (Nunnally, 1967). Following Jehn (1994), we used a median split to determine which groups were high-consensus groups and which were low-consensus groups. Only the high-consensus groups were used in the intensity analysis, based on the previous discussion of the relationship between value congruence and value content intensity (recall that value intensity has little or no meaning in groups with low levels of value congruence). After identifying the nine value dimensions, we coded each individual’s response

15

on each item as being characteristic, uncharacteristic, or neutral for the dimension in which it was included. For example, the Outcome Orientation dimension is made up of 5 value items: achievement orientation (+), being demanding (+), having high expectations for performance (+), being results oriented (+), and being calm (-). Each of these five items was scored as either characteristic, neutral, or uncharacteristic for the Outcome Orientation dimension. The sum of the members’ characteristic scores on each dimension was the value content intensity score used in our analysis. Conflict and group outcome measures. Relationship and task conflict were measured by the Intragroup Conflict Scale (Jehn, 1995). This measure includes eight 7-point Likert-type questions. The Cronbach alpha for the scales of relationship and task conflict were .81 and .86, respectively. Objective performance was measured by rating the group’s final reports. Points were awarded for thoroughness of problem identification, accurate analysis, and the final recommendations to the company. The scale was developed by the faculty involved in the project and ranged from 1 to 100, with 100 being the highest score and 23 the lowest. The final report was rated by two independent raters and the average interrater reliability was .94. Members were also asked to respond to the following question regarding their perception of the group performance on a 7-point Likert scale ranging from 1 = “not at all effective” to 7 = “very effective.” Satisfaction was measured by the Kunin faces scale (1955). This 7-point Likert scale asks members to circle the face that most adequately portrays how they feel about working in the group. The faces range from a very sad face to a very happy face. Results 16

Intercorrelations among the variables in the model and their means and standard deviations are shown in Table 1. Regression analyses were used to test the model and are presented in Tables 2 and 3. Hypotheses 1 and 2 were supported: Value congruence decreased both task and relationship conflict. Hypothesis 3 proposes that greater individual visible demographic dissimilarity positively affects relationship conflict. This was found to be true in the case of sex; groups that were more homogeneously male or female had less relationship conflict. The data supported Hypothesis 4, which states that greater individual informational demographic dissimilarity will positively affect task conflict; educational dissimilarity increased task conflict. ----------------------------------------Insert Tables 1, 2 & 3 about here ----------------------------------------The regressions testing Hypotheses 5 to 13 are shown in Table 3.2 Relationship conflict was found to be negatively related to perceived and objective performance and satisfaction, supporting Hypotheses 5, 6, and 7. Task conflict was also negatively related to perceived performance and satisfaction, providing support for Hypotheses 9 and 10. The relationship between task conflict and objective performance was not significant, but in the hypothesized direction (i.e., task conflict was positively related to performance). To address the multicollinearity of the conflict measures (r=.48), we entered each variable one at a time in the regression and received similar (and slightly more significant)

2

It is important to note that our regressions predicting conflict combine group-level value consensus with individual-level demographics. This is appropriate for the demographic individual differences variables, because both task and relationship conflict, our dependent variables, are measured with individual perceptions (Rousseau, 1985). In contrast, in these regressions, group value consensus is an important contextual factor which varies on a group-by-group basis.

17

results when task and relationship conflict were entered independently. In addition, we did Baron and Kenny’s (1986) test of mediation for the conflict variables and found that relationship mediated. First, relationship conflict (mediator) was regressed on the significant independent variables (value congruence, sex) and found to be significant (Beta = -.35, p < .001, Beta = .10, p < .001, respectively). Task conflict (mediator) was also regressed on value congruence and education and found to be significant (Beta = -.17, p < .01, Beta = .09, p < .05). Second, value congruence and education were significantly related to the dependent variables: objective performance (education Beta = .14, p < .001), perceptual performance (value congruence Beta = -.16, p < .001), and satisfaction (value congruence Beta = -.28, p < .001). Sex was not related to the dependent variables. And third, the effect of the value congruence and education became nonsignificant when the conflict variables were included in the regression analyses on performance (Beta = .08, p = ns, Beta = .04, p = ns, respectively) and satisfaction (Beta = .02, p = ns). Thus, the mediating role of conflict on performance and satisfaction was confirmed in the case of (1) task conflict mediating education dissimilarity and performance, (2) relationship conflict mediating value congruence and perceptual performance, and (3) relationship conflict mediating value congruence and satisfaction. Based on Baron and Kenny’s test of mediation, relationship conflict did not mediate the relationship between sex dissimilarity and the dependent variables. There was some support for H11, which predicted that the intensity of innovativeness, detail orientation, outcome orientation, reward orientation, team orientation, and decisiveness are positively related to performance. Groups that valued outcome orientations and detail orientations had high levels of actual performance, supporting H11. The values of decisiveness, aggressiveness, and stability were 18

positively related to perceived performance (H12) and supportiveness and decisiveness values were positively related to individual satisfaction (H13). Two unexpected results were that outcome orientation was negatively related to perceived performance and a team orientation was negatively related to satisfaction.

Discussion The results of this study indicate that visible forms of work group differences (e.g., sex) increase relationship conflict, while differences regarding informational demographics (e.g., education) increase task-focused conflict. This is one of the first empirical studies to examine different forms of individual demographic dissimilarities (visible and informational) and their impact on perceived intragroup conflict. We also found that the similarity of values among members regarding group processes decreased both task and relationship conflict. Differences in age and nationality did not negatively affect relationship conflict as we predicted. The lack of findings for the differences in age could be due to our sample. The MBA workgroups were comprised of individuals who were for the most part, of roughly the same age. Thus, we would expect that in samples in which there is great heterogeneity in age there would be more relationship conflict. The same argument applies to the nationality variable. As seen in Table 1, the mean for nationality dissimilarity is .28 and the standard deviation is .34, suggesting that the workgroups we looked at were relatively homogeneous in terms of nationality. Once again, we would expect that the results might be different in samples where there is a higher level of nationality diversity. 19

We predicted that the impact of value congruence on conflict would have mixed effects on group performance since past research has demonstrated that while relationship conflict is negatively related to performance, moderate amounts of task conflict can enhance group performance (Jehn, 1995). In our study, relationship conflict decreased satisfaction as well as performance, and while task conflict increased members’ dissatisfaction with the group, it did not influence objective performance. The difference in our findings may be a function of the type of task our groups were performing. Jehn (1995) found task conflict to be productive in groups performing nonroutine tasks but destructive in groups performing routine tasks. Future research needs to examine the aspects of the task that influence the effect of task conflict. We also found that the content of the values influences the performance of the group. While individuals who valued decisiveness were happier and thought their group performed better, these value dimensions were not linked to objective performance; and groups with strong outcome-oriented individuals performed better objectively, but did not believe that they had done so. Interestingly, values that group members intuitively believe will increase performance actually do not increase performance. Groups that valued achievement orientation, being results oriented, being demanding and having high expectations (all components of the outcome dimension) were the highest performers. A detail orientation also increased actual performance even though members did not realize it. Contrary to our predictions, most groups with a team orientation were dissatisfied. This may be explained through the process loss (Steiner, 1972) associated with working in teams such as conflict, coordination problems, and time and scheduling constraints which members indicated led to frustration. However, individuals believed that they had to de-emphasize outcomes and stress decisiveness, 20

stability, and aggressiveness while bridling conflict, especially relationship conflict, in order to perform well, all of which concern group processes. In part, this may reflect the availability of feedback: Chances to evaluate group processes come regularly, but content-related feedback before the termination of the project was rare, and did not formally occur until after our data were collected. It also may reflect our respondents’ implicit preferences for seeing themselves as aggressive, stable, and decisive--not an unreasonable supposition, given our sample. Overall, it is striking to note how slight the correlation is between objective and perceived performance and how strong it is between perceived performance and satisfaction (Table 1). Apparently, individuals’ perceptions of their performance are strongly related to their satisfaction with the group experience, and because of this, individuals are poor judges of their objective performance. Again, future research in contexts with different tasks may illuminate how specific value orientations influence group outcomes: presumably, the impact of different value orientations vary across group task environments. Limitations and Future Research One of the limitations of this study is that the sample was composed of MBA workgroups rather than employee workgroups in an organization. The consultation problems that the workgroups faced were actual organizational issues provided to them by companies. The MBA workgroups worked on site in the organizations whose problems they were trying to solve. Despite this, however, there still remain some differences between MBA workgroups and organizational employees. First of all, the students did not have access to the resources and information that hired consultants have. Secondly, the reward for doing a good job (a grade) may not carry the same effects as a reward in an organization (a 21

job and pay). Finally, the workgroups in our study faced a situation that was clearly temporary and they might never have to face the individuals in their group again; therefore, there may not have been much incentive for individuals to work out their problems with other group members. Another limitation of the study is the common method bias that is present in some of the analyses. While some of the measures (demographic information and objective performance) were gathered from sources other than individual perceptions, values, conflict, satisfaction, and perceived performance were all self-report measures. Value preferences were measured via the Q-sort in a different time period than the other measures, so it is unlikely that any bias exists between this measure and the survey responses. Since the conflict measures, satisfaction, and perceived performance were collected in the same instrument there may be a self-report bias since an individual who is unhappy may be more likely to report higher levels of conflict and lower levels of satisfaction. However, this does not explain the influence of demographic difference and value congruence on conflict. Another important limitation points to possibilities for future research. In particular, applying findings from our sample of MBA student work groups over the course of a semester to work groups in business organizations should be approached cautiously. Because this study was a quasi-experiment, we were unable to control conditions as thoroughly as in a laboratory study. However, our approach can be seen as a reasonable compromise between the controlled artificiality of a lab study and the constantly changing chaos of work groups in business organizations. Nevertheless, we hope to apply a refined version of this paper’s model to future research on work groups in business organizations. Future research in this area should focus on a number of different things. First, this model should be tested on different samples of workgroups that have different levels of diversity to determine 22

what the effects of demographically different individuals are on various outcomes. It is also imperative that more comprehensive demographic measures be collected, such as race, level of education achieved, and organization tenure. The relationship between values and demographic characteristics needs to be explored more fully. Demographic characteristics have been used as indicators of values in most demographic research but the accuracy of this assumption has not been fully tested. In addition, the issue may not really be that demographic characteristics indicate certain values but that people stereotype others. Stereotypes classify individuals into groups and imply certain behaviors and values. The discrepancy between who an individual feels he/she is and how others see him/her may have more of an effect than the level of demographic dissimilarity that occurs between individuals in groups . More research needs to be done to untangle out these relationships. Summary The implications of these findings for managers and group leaders are quite important. When managers have control over group formation they can place together members by assessing the range of values and demographics in the organization to create high performing teams. If the group leader does not have control over group composition, he/she will still be able to predict the type of conflict that is most likely to occur by identifying the values and demographic diversity within the group and will be able to more adequately manage the group process. This could take the form of proactive training which would educate managers, team leaders and members on the benefits and detriments of diversity and conflict within organizational groups. In the interests of promoting diversity, teams are often formed such that they include many 23

people who are different from one another without taking into account some of the problems that may arise. The results of this study have clearly shown that differences contribute to conflict and that conflict often has negative repercussions. Haphazard development of diverse teams without provisions for training and conflict resolution mechanisms may end in unproductive and inefficient use of resources.

24

References Alagna, S.W., Reddy, R.M., and Collins, D. (1982). Perceptions of Functioning in Mixed-sex and Male Medical Training Groups, Journal of Medical Education, 57, 801-803. Amason, Alan C. (1996).Distinguishing the effects of functional and dysfunctional conflict on strategic decision making: Resolving a paradox for top management teams. Academy of Management Journal, 39:123 - 148. Amason, Alan C. & Schweiger, D.M. (1994). Resolving the paradox of conflict, strategic decision making, and organizational performance. International Journal of Conflict Management, 5: 239 253. Ancona, D.G. (1990), Outward bound: Strategies for Team Survival in an Organization, Academy of Management Journal, 33: 334-65. Ancona, D.G., and Caldwell, D.F. (1992), Demography and Design: Predictors of New Product Team Performance, Organization Science, 3: 321-341. Bandura, A. (1982), Self-efficacy mechanism in human agency. American Psychologist, 37: 122-147. Bantel, K.A. and Jackson, S.E. (1989), Top Management and Innovations in Banking: Does the Composition of the Top Team Make a Difference? Strategic Management Journal, 10: 107124. Baron, R.M. & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51: 1173-1182. Brewer, H.B. (1981). Ethnocentrism and its role in interpersonal trust. In M. Brewer & B. Collins 25

(Eds.) Scientific Inquiry and the Social Science, San Francisco: Jossey-Bass. Chatman, J. A. & Jehn, K. (1994). Assessing the relationship between industry characteristics and organizational culture: How different can you be? Academy of Management Journal, 37: 522553. Dienesch, R.M. and Liden, R.C. (1986), Leader-Member Exchange Model of Leadership: A Critique and Further Development, Academy of Management Review, 11: 618-634. Enz, C. (1988). The role of value congruity in intraorganizational power. Administrative Science Quarterly, 35: 504- 529. Enz, Cathy A. and Charles R. Schwenk. (1991). The performance edge: Strategic and value dissensus. Employee Responsibilities and Rights Journal, 4: 75-85. Jackson, S.E (1992). Team composition in organizations. In S. Worchel, W.Wood, & J. Simpson (eds.), Group process and productivity. London: Sage publications. Janis, I.L. (1982), Groupthink (2nd ed.), Boston, MA: Houghton-Mifflin. Jehn, K. (1994). Enhancing effectiveness: An investigation of advantages and disadvantages of valuebased intragroup conflict. The International Journal of Conflict Management, 5: 223-238. Jehn, K. (1995). A multimethod examination of the benefits and detriments of intragroup conflict. Administrative Science Quarterly, 40, 256-282. Jehn, K. (1997). A qualitative analysis of conflict types and dimensions in organizational groups. Forthcoming at Administrative Science Quarterly. Jehn, K. & Shah, P. (1997). Interpersonal relationship and task performance: An examination of mediating processes in friendship and acquaintance groups. Forthcoming at Journal of 26

Personality and Social Psychology. Kanter, R. M. (1977). Men and women of the corporation. New York: Basic Books. Kent, R.N. and McGrath, J.E. (1969), Task and Group Characteristics as Factors Influencing Group Performance, Journal of Experimental Social Psychology, 5, 429-440. McPherson, J. M. & Smith-Lovin, L. (1987). Homophily in voluntary organizations, status distance and the composition of face-to-face groups. American Journal of Sociology, 52: 370-379. O'Reilly, C. A., Chatman, J. & Caldwell, D.F.. (1991). People and organizational culture: A profile comparison approach to assessing person-organization fit. Academy of Management Journal, 34: 487-516. Pelled, L.H. (1996), Demographic Diversity, Conflict, and Work Group Outcomes: An Intervening Process Theory, Organization Science, forthcoming. Rokeach, M. (1973). The nature of human values. New York: The Free Press. Rousseau, D. (1985). Issues of level in organization research: Multi-level and cross-level perspectives. In L.L. Cummings & B.M. Staw (Eds.), Research in organizational behavior (Vol. 7, pp. 137), Greenwich, CT: JAI Press. Schein, Edgar H. (1985). Organizational culture and leadership. San Francisco: Jossey-Bass. Shah, P. & Jehn, K. (1993). "Do friends perform better than acquaintances?: The interaction of friendship, conflict and task," Group Decision and Negotiation, 2, 149-165. Steiner, I.D. (1972). Group process and productivity. San Diego, CA: Academic Press. Tsui, A.S., Egan, T.D., and Xin, K.R. (1995), Diversity in Organizations: Lessons from Demography Research. In M.M. Chemers, S. Oskamp, M.A. Costango (Eds.) Diversity in Organizations: 27

New Perspectives for a Changing Workplace, Thousand Oaks, CA: Sage Publications, Ltd. Tsui, A.S., Xin, K.R., and Egan, T.D. (1995), Relational Demography: The Missing Link in Vertical Dyad Linkage. In S. Jackson, M. Ruderman, & W. Tornow (Eds.) Work Team Dynamics and Productivity in the Context of Diversity, American Psychological Association. Tsui, A.S. and Egan, T.D. (1994), Performance Implications of Relational Demography in Vertical Dyads, Paper presented at the National Academy of Management meeting, Dallas, Texas. Tsui, A.S., Egan, T., and O’Reilly, C.A.III. (1992), Being Different: Relational Demography and Organizational Attachment, Administrative Science Quarterly, 37: 549-79. Watson, W.E., Kumar, K., and Michaelsen, L.K. (1993), Cultural Diversity’s Impact on Interaction Process and Performance: Comparing Homogeneous and Diverse Task Groups, Academy of Management Journal, 36: 590-602. Zenger, T.R. & Lawrence, B.R. (1989). Organizational demography: The differential effects of age and tenure distributions on technical communication. Academy of Management Journal, 32: 353376.

28

Karen A. Jehn Clint Chadwick Sherry M.B. Thatcher 2000 Steinberg Hall - Dietrich Hall University of Pennsylvania Philadelphia, PA 19104-6370 tel. 215 898-0525 fax 215 898-0401 [email protected]

Dr. Karen A. Jehn is Associate Professor of Management at the Wharton School at the University of Pennsylvania. Her current research focuses on intragroup and intergroup conflict, deception in organizations, and bicultural teams.

Clint Chadwick is a doctoral candidate at the Wharton School at the University of Pennsylvania. His research interests include the strategic 29

management of human resources, downsizing, restructuring, and strategic renewal in organizations, and the performance impact of values in groups and organizations.

Sherry M.B. Thatcher is a doctoral candidate at the Wharton School at the University of Pennsylvania. Her current research focuses on group diversity, and the fit between individuals’ identities and others’ categorizations of those individuals. She is also interested in prejudice and stereotypes in multinational corporations.

30

Figure 1 Model of Value Congruence, Individual Demographic Dissimilarity, Conflict and Group Outcomes

Value Congruence

Value Content Intensity* Task Conflict

Actual Performance Perceptual Performance

Individual Demographic Dissimilarity

Relationship Conflict

Visible Informational

* Assumes high value congruence. See footnote 1.

31

Member Satisfaction

Table 1 Means, Standard Deviations, and Intercorrelations * 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

Variable 1. Value Congruence 2. Sex Dissimilarity 3. Age Dissimilarity 4. Nationality Dissimilarity 5. Education Dissimilarity 6. Innovation 7. Stability 8. Detail Oriented 9. Outcome Oriented 10. Aggressive 11. Reward 12. Supportive 13. Team Oriented 14. Decisiveness 15. Task Conflict 16. Relationship Conflict 17. Objective Performance 18. Perceived Performance 19. Satisfaction

.15 .10 -.21 -.15 .10 -.14 -.13 -.10 .21 -.02 .21 .17 .04 -.15 -.23 .04 -.17 -.30

.02 -.06 .01 .03 -.07 .05 -.07 -.10 -.09 .08 .13 .03 .02 -.07 .01 -.04 .07

-.11 -.05 .14 -.05 .04 .01 .01 -.07 .05 -.15 -.08 -.14 .05 .07 .11 .06 -.05 .03 .07 .09 .07 .03 -.04 .08 -.06 .06 .02 -.06 .03 -.02 .05

.01 .06 -.45 .08 -.08 -.06 -.01 .08 .12 -.13 .14 -.11 -.09 .09 .08 -.05 -.18 -.07 .09 -.02 .13 .16 -.01 -.01 -.14 .07 -.16

.17 .03 -.10 .12 -.15 -.11 .11 -.07 -.15 -.01 .18 .15

.18 .09 -.12 -.29 -.14 -.05 -.17 -.11 .05 .00 .11

.33 -.07 -.06 -.33 -.21 -.10 -.11 -.16 -.05 .24 -.12 -.06 -.01 -.07 -.16 -.21 .16 .07 -.27 .08 -.22 .17 -.08 -.30 -.24 .14 .04 .08 -.16 -.04 -.01 .07 .08 .18 .28 .10 .12 .04 -.28 .28

.14 .06 .48 .09 -.11 -.15 .13 -.13 -.38 -.06 .14 -.19 -.50 -.07

.64

Mean .59 .52 2.86 .28 .56 .26 .17 .35 .28 .30 .06 .38 .27 4.45 5.35 87.36 3.85 2.44 s.d. .18 .26 2.01 .34 .26 .18 .17 .31 .20 .19 .13 .18 .22 1.51 1.66 6.27 .83 1.27 __________________________________________________________________________________________________________________________________ * all correlations with absolute values above .10 are significant at p