Page 86 - Discrimination at Work The Psychological and Organizational Bases

RIORDAN, SCHAFFER, STEWART
56
different from themselves on race and gender (and other attributes) may be
adding to the difficulty of interaction for these groups" (pp. 414-415). Fi­
nally, Brickson (2000), in her review of identity orientations, suggested that
the activation of one's personal and social identities brings about mecha­
nisms that can lead to prejudice and discrimination. Together, these works
support the idea that employees who are demographically dissimilar to
other workgroup members will feel the effects of outgroup bias and dis­
crimination.
Previous research has also shown that perceived discrimination is an­
tecedent to a number of important work-related outcomes. Many of these
outcomes have been used as dependent variables in relational demogra­
phy research. For example, studies have shown that employees' beliefs
about discrimination, whether they are accurate or not, can affect psycho­
logical and physical health and adjustment (Barak, Cherin, & Berkman,
1998; Eisenberger, Fasolo, & Davis-LaMastro, 1990; Ensher, Grant-Vallone,
& Donaldson, 2001); job involvement, career satisfaction, and career
prospects (e.g., Foley & Kidder, 2002; Perry, Hendricks, & Broadbent, 2000;
Sanchez & Brock, 1996; Shaffer, Joplin, Bell, Lau, & Oguz, 2000; Valentine,
Silver, & Twigg, 1999); work conflict, lower feelings of power, and job pres­
tige (Gutek, Cohen, & Tsui, 1996); turnover intentions (Shaffer et al., 2000);
and organizational citizenship behaviors (Ensher et al., 2001). In short,
discrimination-type outcomes may act as mediators between demographic
dissimilarity and some of the more traditional outcomes of relational de­
mography. Future research should explore these linkages.

Contextual and Personal Factors as Moderators

As depicted in Fig. 3.1, future research on relational demography needs
to explore the role of moderators or the conditions under which the re­
lationships between demographic dissimilarity and the outcomes are the
strongest/weakest. Earlier studies have not supported the moderating ef­
fects of minority status (Jackson et al., 1991; Kirchmeyer, 1995). Recently,
Flynn et al. (2001) found that dissimilarity in citizenship, race, and gender
was related to individuals forming more negative impressions of others.
However, this relationship was mitigated or buffered when dissimilar in­
dividuals were more extraverted or when they were higher self-monitors.
In this case, extraversion and self-monitoring were contextual variables
(moderators) in the design.
The use of theoretically relevant moderators appears to provide ad­
ditional information to the interpretation of observed results, and future
studies in relational demography should therefore continue to incorporate
such variables as part of their overall research designs. For example, as