Page 375 - Discrimination at Work The Psychological and Organizational Bases

PAETZOLP
342
DISPARATE IMPACT AND CONTEMPORARY PRACTICES
Intentional discrimination is not the only form of discrimination that has
contributed to differential outcomes for protected groups. The disparate
impact model, which requires no proof of intent to discriminate, is used to
challenge employment practices or policies that are facially neutral but that
have an adverse impact on a protected group. In this section, I first focus on
statistical issues that reflect the individual nature of the disparate impact
model (counter to common conceptions among I/O and HR researchers).
Second, I briefly address growing concerns about the focus on "fit" in
organizational research and potential conflicts under the disparate impact
model of discrimination.

Multiple Screens and Heterogeneity

The disparate impact model has played a significant role in eliminating dis­
crimination that appears to be unintentional, at least on its face. Although
theoretical justifications for this model vary, making it a rather controver­
sial model at law (Caldwell, 1985; Chamallas, 1983; Fiss, 1971; Paetzold &
Willborn, 1996; Perry, 1977; Rutherglen, 1987), it is well-established and
well-recognized among the I/O and HR communities. Causation is cen­
tral to this model. To establish a prima facie case of disparate impact, a
plaintiff must demonstrate that that an employer "uses a particular em­
ployment practice that causes a disparate impact on the basis of race, color,
religion, sex, or national origin" (Title VII, 1993, emphasis added). [Be­
cause this language does not appear in other antidiscrimination statutes,
such as the Age Discrimination in Employment Act (1985), its existence
for proving other types of discrimination is unclear (e.g., Hazen v. Biggins,
1993).] If the plaintiff establishes the prima facie case—typically by using
either the four-fifths rule or other statistical evidence—then the burden of
proof shifts to the employer to prove business necessity or job-relatedness
(Griggs v. Duke Power Company, 1971).
The causation element in disparate impact analysis is a complex one, as
has been argued by Paetzold & Willborn (1996,2002). Obviously, the model
does not require proof that the cause of the discrimination is the result of
bad motives; in addition, it does not account for the fact that there are
always multiple causes why a particular employment screen would tend
to have an adverse impact on a protected group. Among others, social
and/or psychological causes contribute to all disparate impacts. For ex­
ample, word-of-mouth recruiting may tend to replicate a predominantly
White work force because of "similar-to-me" biases in who receives the
"word of mouth" (see, e.g., Rand & Wexley, 1975, for a discussion of the