Exploring the black box: an analysis of work group diversity, conflict, and performance

Administrative Science Quarterly,  March, 1999  by Lisa Hope Pelled,  Kathleen M. Eisenhardt,  Katherine R. Xin

• E-mail
• Print
• Link

If we cannot now end our differences, at least we can help make the world safe for diversity.

- John F. Kennedy

In the past decade, demographic diversity has become one of the foremost topics of interest to managers and management scholars. The term demographic diversity refers to the degree to which a unit (e.g., a work group or organization) is heterogeneous with respect to demographic attributes. Attributes classified as demographic generally include "immutable characteristics such as age, gender, and ethnicity; attributes that describe individuals' relationships with organizations, such as organizational tenure or functional area; and attributes that identify individuals' positions within society, such as marital status" (Lawrence, 1997: 11). The heightened concern with demographic diversity (hereafter referred to simply as diversity) stems not only from the growing presence of women and minorities in the work force (Buhler, 1997) but also from modern organizational strategies that require more interaction among employees of different functional backgrounds (e.g., Dean and Snell, 1991). One of the most significant bodies of research to arise from this trend is a stream of field studies linking group composition to cognitive task performance - i.e., performance on tasks that involve generating plans or creative ideas, solving problems, or making decisions. The impact of diversity on cognitive task performance has been examined in studies of top management teams (e.g., Bantel and Jackson, 1989; Murray, 1989; Eisenhardt and Schoonhoven, 1990) and lower-level work groups (e.g., Kent and McGrath, 1969; Murnighan and Conlon, 1991).

Despite this spotlight on diversity in work groups, there is more to be done. Investigations of diversity and work group performance have largely been what Lawrence (1997) referred to as "black box" studies, which do not measure intervening process variables. Further, the effects on performance are still unclear. Some studies (e.g., Bantel and Jackson, 1989) have linked diversity to favorable performance on cognitive tasks, and some (e.g., Murnighan and Conlon, 1991) have linked it to unfavorable performance on such tasks. Others (e.g., Watson, Kumar, and Michaelson, 1993) have shown that group diversity both enhances and diminishes cognitive task performance. To capture fully the complex relationship between work group diversity and performance, we need more sophisticated theories and empirical work incorporating intervening variables and multiple types of diversity. The objective of the current investigation is to begin to meet these needs, offering an intervening process theory - one that attempts to untangle the complicated set of relationships among five types of diversity and performance - and providing a test of that theory.

Two prior studies that have empirically assessed whether process variables intervene between group diversity and performance are particularly important to our efforts. Ancona and Caldwell (1992) examined the intervening role of internal task process (i.e., the setting of goals and priorities) and external communication. Later, Smith et al.'s (1994) top management team study looked at three potential intervening variables (social integration, informality of communication, and communication frequency). The authors of both studies discovered that the process variables they measured did not fully explain the observed effects of diversity on performance, and both then suggested that the mediating effect of conflict should be assessed in future research. Hence, in the model we propose and test here, conflict plays an intervening role.

The model proposes that work group diversity indirectly affects cognitive task performance through two kinds of conflict: intragroup task conflict and intragroup emotional conflict. Task conflict is a condition in which group members disagree about task issues, including goals, key decision areas, procedures, and the appropriate choice for action, and emotional conflict is a condition in which group members have interpersonal clashes characterized by anger, frustration, and other negative feelings (Jehn, 1994; Eisenhardt, Kahwajy, and Bourgeois, 1997a).(1) We suggest that job-related types of diversity largely drive task conflict. In contrast, emotional conflict is shaped by a complex web of diversity types that increase emotional conflict based on stereotyping and decrease emotional conflict based on social comparison. Task routineness and group longevity moderate these diversity-conflict relationships. Each type of diversity indirectly affects performance via its relationship with conflict: task conflict tends to enhance performance, while emotional conflict tends to diminish performance. Thus, we offer a model that postulates that the black box between diversity and performance contains a more elaborate set of relationships than previously thought.

THEORETICAL BACKGROUND AND HYPOTHESES

Link between Diversity and Task Conflict