The Association Between Ethnic Congruence in the Supervisorâ•fiSubordinate Dyad and Subordinate Organizational Position and Salary

We used a large field sample (N = 32,854) to test popular press claims and propositions from relational demography theory that the ethnic congruence of the supervisor–subordinate dyad is associated with positive employee outcomes. Results indicate that ethnic congruence was positively associated with subordinates’ level in the organization and salary. These relationships were moderated by type of compensation system, such that the effects were stronger in merit-based than in tenure-based systems. In addition, the extent of the relationship between supervisor–subordinate congruence differed by ethnic group, supporting previous research that suggests examining unique outcomes by ethnicity. However, although statistically significant, effect sizes were generally smaller than anticipated and implied in the popular press. Practical implications of these findings are discussed. The popular press has put forth a consistent argument for the value of ethnic diversity in organizations (Fernandez, 1995; Martinez, 1995; Varian, 2005). Various researchers (e.g. McLeod & Lobel, 1996; Richard, 2000; Watson & Kumar, 1993) have sought to validate these assumptions, while others have prescribed models to leverage diversity beneficial to organizations and individuals (Cox, 1991; Cox, Lobel & McLeod, 1991). Although “diversity” may be used to refer to a multitude of individual and organizational characteristics, we are concerned here primarily with ethnic diversity or the racial complexity of an organization. Diversity research in organizations has proposed that working with a same ethnicity supervisor will provide professional and personal support and motivation, bolstering a minority employee’s level of efficacy and success within an organization while at the same time reducing potentially negative or neutral mixed-ethnicity-only situations (Avery, 2003; Tsui & O’Reilly, 1989). 552 Avey, West, & Crossley in J . Occ. Org. Psych. 81 (2008) Contrarily, diversity research has also proposed the disadvantage of incongruence of ethnicity in the supervisor–subordinate dyad where minority employees have reported less access to career-enhancing networks (Ibarra, 1995) and overall less satisfaction (Greenhaus, Parasuraman, & Wormley, 1990). For instance, examining ethnic incongruence, Jeanquart-Barone (1993, 1996) argued that Black employees with White supervisors report less supervisor support and fewer developmental opportunities, leading to potentially lower employee level attainment within an organization. However, the majority of studies conducted to investigate the potential value of diversity and supervisor–subordinate ethnic congruence have focused primarily on attitudinal and perceptual outcomes, with performance being one of the few behavioral outcomes occasionally incorporated (Ibarra, 1995; James, 2000; Jeanquart-Barone, 1993, 1996; Kraiger & Ford, 1985; McLeod & Lobel, 1996; Perkins & Thomas, 2000; Pulakos, White, Oppler, & Borman, 1989; Richard, 2000; Tsui & O’Reilly, 1989; Watson & Kumar, 1993). Although extent research in organizational behavior has shown employee attitudes towards supervisors and the organization as a whole may be meaningful, from the employee perspective more objective and practical outcomes are likely to also to be of interest. However, there has been a lack of actual results showing specifically that the assumed impact on promotion and salary is a reality. While beliefs regarding these phenomena are no doubt important and influential in many ways, it is vital to understand the actual impact on organizational outcomes if we are to fully grasp the importance of congruent ethnicity supervisor–subordinate relationships within organizations. Addressing these issues will begin to shed light on questions such as whether organizations may need to re-evaluate their performance management and promotion systems in a diverse environment or, less drastically, make an effort to manage the impressions and beliefs that employees hold regarding the potential for advancement. Therefore, the purpose of the present study was to examine whether supervisor–subordinate ethnic congruence is in fact associated with practically important subordinate outcomes such as salary and existing level within the organization, as well as the magnitude of these relationships. In addition, to better understand these potential relationships, we examine the extent to which the relationships differ across both compensations systems and specific ethnicity. In this study, we consider the supervisor as the immediate manager rather than any manager hierarchically above the employee in terms of organizational level. A previous study examining rater bias by Pulakos and colleagues (1989) using nearly 20-year-old data partially addressed this question in an analysis of Army participant performance ratings. Specifically, they examined the effects of rater and ratee race on subjective performance ratings of first-term Army enlisted personnel. Using both cross-sectional and repeated measures analyses, they found that ratees were rated slightly higher by raters in same-race dyads and noted “the overwhelming finding was that irrespective of whether between-subjects or repeated measures were used, the proportions of rating variance accounted for by race and gender were extremely small.” We extend this research in the current study by (1) conducting these analyses with more recent data; (2) conducting in the private sector; (3) specifically focusing on practical and objectively measured outcomes of salary and level; and (4) using a structural moderator in the form of the organizational compensation system. Theory and hypotheses The primary proposition underlying the literature applicable to this study is that there are benefits for those employees who are ethnically similar to their supervisor. Two theoethniC CongruenCe And subordinAte pos it ion And sAlAry 553 retical frameworks have primarily been used to explain this relational demography effect; the similarity-attraction paradigm and social identity theory. The similarity-attraction paradigm first proposed by Byrne (1971) argues that the extent to which an individual perceives another individual to be similar to themselves, they will be seen as more attractive. Although this judgment may fluctuate over time, it influences the level of attraction between individuals. Avery (2003, p. 672) used this framework to describe his findings when he argued “High perceptions of (ethnic) similarity tend to elicit favorable responses such as interpersonal attraction, perceptions of procedural fairness, and increased job satisfaction.” Extending the similarity-attraction paradigm into a more explanatory framework, Tsui, Egan, and O’Reilly (1992, p. 551) argue “The similarity-attraction hypothesis assumes interaction among individuals. Although interpersonal interaction is thus a necessary condition for social integration to occur, individuals can express preferences for a group even without social interaction.” They conclude that “the similarity-attraction paradigm may not account for all the reported demographic effects, especially when actual interaction among the participants is unlikely.” Based on these suggestions, social identity theory (Tajfel, 1982) and more specifically self-categorization theory (Turner, 1982) have been proposed as a potentially more comprehensive explanatory framework. According to self-categorization theory, people classify themselves on the basis of surface-level diversity (e.g. demographic attributes such as ethnicity and gender) to form “in-groups” and “out-groups.” In other words, “people use demographic differences, particularly those that are visible, to categorize one another” (Chapman & Spataro, 2005, p. 322). Brewer and Lui (1989) note an exceptionally strong tendency to classify based on visible characteristics beyond other potentially assumed characteristics, given that visible characteristics are considered to be more known than those characteristics that must be assumed (e.g. moral character). Furthermore, people subsequently bias in favor of similar in-group members and bias against out-group members (Hewstone, Rubin, & Willis, 2002). This effect is increased as individuals denigrate out-group members and perceive them to be less trustworthy, honest, cooperative, and effective than in-group members. These assertions were supported by Tajfel (1982) who conducted 30 field studies validating out-group and in-group biases. Therefore, as applied to a manager–subordinate dyad within an organizational setting, social identity theory would suggest that the extent to which the manager perceives the employee, based on visible characteristics, to be more like himself/herself, the employee is more likely to be classified as an “in-group” member (receiving the accompanying positive bias) and less likely to be an “out-group” member (receiving the accompanying denigrating bias). Given that managers are often in positions to determine employee salary and level, this positive or denigrating bias is likely to influence these outcomes. This is discussed below specifically in the context of ethnic congruence or incongruence in the dyad. Effects of ethnic congruence Research using self-categorization theory and focusing on ethnic congruence in the supervisor–subordinate dyadic relationship has also been analyzed considering social networks, job attitudes, and subjective performance outcomes and found support for the general relational demography hypotheses. For example, Ibarra (1995) demonstrated 554 Avey, West, & Crossley in J . Occ. Org. Psych. 81 (2008) that minority employees had fewer intimate social networks in the organizational setting and perceived less access to career benefits from those networks. This phenomenon is consistent with self-categorization theory, whereas individuals who are members of a supervisor’s “in-group” are likely to experience more social interaction with the supervisor, thus potentially leading to expanded social networks in the organization. It is important to note that both Ibarra (1995) and Jeanquart-Barone (1993, 1996) demonstrate a belief of minority employees of less developmental opportunities and less benefit from identical social networks than non-minority counterparts. This belief is further enhanced for minority employees experiencing no ethnic congruence in their supervisor–subordinate dyads, as when minority employees see representation of their ethnicity at all levels of the organization they tend to perceive both more opportunities to advance (Kanter, 1977) and improved promotion rates (Konrad & Pfeffer, 1991; James, 2000). Considered in a self-categorization theory framework, these results may generalize to White employees with minority supervisors. Self-categorization does not explicitly account for ethnic majority versus minority effects but rather argues that all individuals classify other individuals into more in-group or out-group status according to observable stimuli. Thus, according to self-categorization theory, a White subordinate with a minority supervisor is likely to experience these perceptions similar to a minority subordinate with a White supervisor. A key question arises with regard to whether these beliefs and assumptions are supported by actual workplace behavior and not just the theories underlying them. Following from earlier discussion on self-categorization theory, this framework suggests that White supervisors are likely consciously or subconsciously demonstrating denigrating bias towards minority subordinates in favor of other in-group White subordinates. This positive bias for in-group members is likely to manifest in more positive performance evaluations, favorable work assignments, increased recognition, and an improved likelihood of “selling” an in-group subordinate during discussions of promotion or pay decisions. For example, a meta-analysis (Kraiger & Ford, 1985) found that ethnic congruence with the manager was associated with higher performance evaluations. As previously noted and consistent with Kraiger and Ford (1985) and Jeanquart-Barone (1993, 1996) found that ethnically different leader–follower dyads yielded lower trust and less supervisory support than same ethnicity dyads. In addition, Greenhaus et al. (1990) found that minority status in organizations (measured by Black and White) was negatively related to supervisor rating of promotability and positively related to early career plateau. They note that this can be attributed to “subgroup membership” and may include effects on salary increases, career advancement, and psychosocial support (p. 65). Finally, Pulakos and colleagues (1989) also found that demographic similarity was associated with more supporting relationships. If this positive bias for in-group members and denigration for out-group members is occurring, it should have significant implications for workplace outcomes of salary and organizational level specifically as a consequence of the supervisor–subordinate relationship. Another aspect to consider is that the beliefs described above may act as somewhat of a self-fulfilling prophecy. Specifically, these beliefs may be powerful predictors of performance, as higher levels of personal efficacy have been associated with higher performance and effort (Bandura, 1997; Stajkovic & Luthans, 1998). Following from this line of research, minority employees who do not perceive an equal opportunity to advance (i.e. lower efficacy towards career advancement) may become less likely to continue to put ethniC CongruenCe And subordinAte pos it ion And sAlAry 555 forth the effort required to warrant advancement, viewing the goal as unobtainable and not seeing a link between performance and outcomes for themselves. Such a situation may result in minority employees effectively bolstering any negatively biased attitudes on the part of White supervisors, and only serving to perpetuate the problem. Similar to the previous scenario, behavioral reactions to strongly held beliefs about ethnic congruence in organizations, if occurring, should have deleterious effects on the actual promotion and salary decisions that are made with regard to minority employees in non-congruent supervisor–subordinate relationships. The preceding discussion and findings suggest that ethnic congruence in the supervisor–subordinate dyad results in positive outcomes such as more interaction between supervisor and subordinates, higher level of subordinate job efficacy, lower levels of absenteeism, higher levels of organizational commitment, and higher performance ratings. Building on to this research, it is generally expected that higher salary and level within the organization should follow. Thus, we present the following hypotheses: H1: Employees with same ethnicity as their supervisors will have a higher level in the organization than employees with a supervisor of different ethnicity. H2: Employees with same ethnicity as their supervisors will have higher salary than employees with a supervisor of different ethnicity. Effects of context After reviewing the multitude of studies regarding relational demography, Avery (2003) concluded that few moderators have systematically been considered in the relational demography research stream. Contextual moderators, whereby situational factors may influence the size and frequency of the effects of relational demography dimensions on employee outcomes, can be important in explaining what Sutton and Staw (1995) call “micro processes” that encompass true theory. As previous research has suggested, employee attitudes (e.g. job satisfaction) may be influenced by ethnic congruence with the supervisor (Avery, 2003). This phenomenon would likely generalize to employees in both meritand tenure-based compensation systems. However, it is quite possible that the effects on organization outcomes such as level and salary may be suffocated specifically by a tenure-based compensation system. For example, exclusively tenure-based compensation systems are those that promote based on years of service with the company. When this system is practiced, supervisors have little or no direct effect on the promotion of subordinates (short of firing them). Likewise, when salary is exclusively determined by tenure, the effects of supervisor-rated performance are not influential in determining these outcomes. Exclusively seniority-based pay structures hinder growth based on motivation or performance. Thus, the effects of ethnic congruence with one’s direct supervisor are less relevant for level and salary outcomes. Further, in a pure tenure-based system, where years of service is the only influential variable for level and salary, one would expect no relationship between ethnic congruence in the supervisor–subordinate dyad and employee outcomes of salary and level. Contrarily, effects of congruence of ethnicity in supervisor–subordinate dyads are anticipated to be more salient in a merit pay system where supervisors may have more autonomy for direct control over the subordinates’ level and salary. Considering Tsui and O’Reilly’s (1989) findings that supervisors prefer subordinates who are more demographically similar, subordinates who are ethnically similar to their supervisors may be more 556 Avey, West, & Crossley in J . Occ. Org. Psych. 81 (2008) appropriate to receive training and mentoring, higher performance ratings, or increased access to other opportunities. In a merit pay context, supervisors have certain levels of direct control over the salary and level of their immediate subordinates in the form of recommendations for promotion and determining wage increase. In these environments, it is expected that outcomes related to ethnic congruence such as social interaction, developmental opportunities, and likeability and mentoring would be less constrained than in tenure-based systems. Therefore, the relationship between ethnic congruence and employee outcomes is likely to be strengthened in a merit-based system. Thus, building on Hypotheses 1 and 2 that ethnic congruence in a supervisor–subordinate dyad will relate to level and salary, we also argue that in a merit-based system these effects will be even more pronounced. H3: The type of pay system will moderate the relation between ethnic congruence and organizational level such that employees with same ethnicity supervisors will have higher level in the organization under merit pay system, but not under tenure pay system. H4: The type of pay system will moderate the relation between ethnic congruence and salary such that employees with same ethnicity supervisors will have higher salary under merit pay system, but not under tenure pay system.