Mobile Edition
Print
Get the Mag
Weekly UpdatesMuch attention is being paid to the topic of diversity in the workplace, as academicians and managers confront the issues emerging from an increasingly diverse American workforce (Thomas and Ely, 1996; Fagenson, 1993). The impact of increased workforce diversity touches virtually all management concerns. As a result, the diversity literature includes a large number of areas of inquiry including, though not limited to, the impact of diversity on communication-oriented outcomes, on cognitive outcomes like creativity of solutions, and on affective reactions like commitment or satisfaction (including turnover and job performance) (Milliken and Martins, 1996).
Broadly considered, the literature falls into two categories: the impact of the organization on diverse groups and the impact of diverse groups on the organization. Research within the former category focuses on the difficulty that organizations have in providing fair treatment to diverse groups with differing needs and on the challenges of managing heterogeneous groups. Reasons for difficulty in integrating workers range from social constraints that preclude integration and acceptance of women and minorities into the workgroup (Fagenson, 1993; Tsui et al., 1992) to the conflict faced by managers when weighing diversity objectives against resource constraints (Barry and Bateman, 1996). The majority of this writing paints a negative portrait of the organization's attempt to handle diversity within its ranks.
The latter category (i.e., the impact of diverse groups on the organization) largely argues for the potential benefits that are posited to derive from diversity in organizations. Many authors have suggested that ethnic and gender diversity can provide organizational benefits ranging from competitive advantages such as marketing strength and the ability to attract human resources at the strategic level (Cox and Blake, 1991) to improved creativity in group decision making (Nemeth, 1986) and improved task performance (Shaw, 1983). However, studies also indicate that gender and culture diversity create barriers to effective interactions that hinder performance. The net result is that research suggests that, without benefit of effective intervention, there should be as little diversity as possible in decision-making groups (Maznevski, 1994), although, over time, higher levels of diversity may contribute to improved decision making (Watson et al., 1991).
An important but ignored topic of study in the research on group diversity is the basis for work group formation. Many organizational groups, such as functional departments, may be experiencing greater gender and culture diversity as the increasing diversity of the workforce brings a more diverse set of workers to organizations. However, the inflow of diverse workers does not necessarily mean that all organizational groups will assemble in a diverse way. With regard to this issue, Blau's (1977) discussion of group heterogeneity and social structure may be instructive. On the one hand, Blau asserts that similarities on one nominal parameter (e.g., race) will promote social associations. On the other hand, he maintains that people will associate not only with members of their own groups but also with members of other groups.
Of these two competing propositions, it appears that the one emphasizing homogeneity has primacy in the formation of many groups. Research on both similarity-attraction theory (Byrne, 1971) and self-categorization theory (Tajfel and Turner, 1986) suggests that people are likely to be attracted more to those who are similar on demographic characteristics such as race, age, and gender. Kandel's (1978) study of the friendships of 2,000 high school students, for example, found that demographic characteristics of race and gender were the primary correlates of interpersonal attraction in dyads. Work by Osbeck et al. (1997) also supported the similarity-attraction hypothesis, showing a particularly strong relationship for minority groups. Thus, as Makower (1995) has suggested, there is evidence to support the notion that self-organized groups will tend to be homogeneous.
Likewise, it seems probable that informal groups within organizations, as well as work groups that are given discretion in their assembly, will tend to be homogeneous. Further, since informal groups established at an earlier point in time may form the basis for social capital in the future by providing information about career opportunities (Ibarra, 1992) and by increasing the probability of preferential treatment for members of the informal network (Cialdini, 1987; Moscovici, 1985), homogeneity within informal groups may contribute to homogeneity at various levels and in various departments within the more formal organization.
A few studies have looked at the similarity-attraction hypothesis, in particular, as a rationale for group homogeneity within organizations. In the marketing literature, both Dwyer (1998) and Smith (1998) found that gender and age/life stage similarity facilitate buyer-seller relationships. Ibarra (1992), in his study of gender differences in network structure, explored preferences in communication, advice, support, influence, and friendship networks in an advertising firm. Ibarra found that both men and women preferred to establish homophilous networks for friendship and support.
Importantly, however, Ibarra found that, while men evidenced a great degree of sex-based homophily for all types of networks studied, women preferred a greater proportion of men for advice and influence ties. This distinction suggests that the similarity-attraction hypothesis may have greater salience for social networks than for task-based networks. It also suggests other factors that might affect alignment, such as the possibility of tangible rewards from interactions with dissimilar others (Chattopadhyay, 1999). As an example, individuals may prefer interaction with higher-status others in order to gain access to valued resources (Lin, 1982), even if those individuals are dissimilar on demographic variables.
Interestingly, extant research does not address preferences for group composition in workplace contexts where social familiarity is not the principal consideration. It also rarely takes place in settings where the antecedent conditions to group formation are clearly known. Rather, what is often observed is an outcome such as a friendship tie (Mehra, Kilduff et al., 1998) and not the actual process of assembly. Likewise, little research is conducted in settings with sufficient gender or cultural diversity for hetergeneous groups to form in any significant number. The homogeneous alignments often seen in real world settings are frequently only observed outcomes, where the potential for formation of heterogeneous groups can not be known. Indeed, much of management's knowledge of group process is culled from research either on homogeneous groups or on groups in which diversity is defined by dimensions other than culture, such as age and job tenure (Larkey, 1996).
This study expands the existing body of knowledge regarding group assembly by empirically examining a subject that has received no such attention: the gender, racial and ancestral composition of self-formed strategic decision-making groups. Specifically, the study investigates the tendency on the part of a larger, diverse group to form into smaller, homogeneous task groups in a setting in which the movement of individual members into smaller groups is observable and in which the diversity of groups prior to alignment is known. The focus is on racial, ancestral, and gender diversity, which are the dominant kinds of diversity cited in the literature on workforce diversity (Harrison et al., 1998). The dimension of ancestral diversity is especially uncommon and, though related to race, is clearly different and possibly more influential as a variable for social alignment. Overall, this study attempts to answer the question: What level of gender, racial and ancestral diversity will occur in self-formed work groups ?
HYPOTHESES
In light of the fairly consistent body of research suggesting that individuals prefer interacting with members of their own kind, it is the position of this study that students will cluster in homogeneous work groups with regard to gender and cultural variables when they are permitted to form their own groups. In contrast, groups developed through random assignment should not be subject to this form of bias and, as a result, should have greater diversity or heterogeneity on these measures. Thus, the diversity occurring in randomly formed groups should provide an unbiased estimate of the heterogeneity possible in a set of groups absent any concerted effort to maximize diversity. Since self-formed groups should tend toward homogeneity, the diversity of groups formed through self-selection ought to be less than that occurring in randomly formed groups. Given that this study seeks to consider gender, racial, and ancestral diversity, three hypotheses are offered:
Hypothesis 1: Self-formed work groups will have less gender diversity than groups formed through random assignment.
Hypothesis 2: Self-formed work groups will have less racial diversity than groups formed through random assignment.
Hypothesis 3: Self-formed work groups will have less ancestral diversity than groups formed through random assignment.
METHODOLOGY
SAMPLE
Participants in the study were 358 seniors enrolled in multiple sections of an undergraduate capstone business strategy course at a large university in New York City. As part of the course requirements, student teams competed in a computer-simulated business strategy game. "Winning" was a meaningful task-based goal because a large percentage of the students' grades (from 30 to 40%) was based on their team's standing in the game at the end of the semester.
FEED