Section Twois more likely to be based on perceptions of their qualifications.
These perceptions are affected by the fact that, on average,
minorities and females have less training and education.^40 As
Roosevelt Thomas has stated: “Companies are worried about
productivity and well aware that minorities and women
represent a disproportionate share of the undertrained and
undereducated.”^41 To the extent that companies generalize
from classes to individuals and make unfounded attributions,
they underestimate the value and contribution that minorities
and females can make. Not surprisingly, the careers of
minorities and females tend to plateau earlier than those of
male nonminorities.
Two caveats on the implications of the diversity literature
seem warranted. One is that some of the empirical literature
recommends that researchers should seek to determine the
optimal level of heterogeneity because there is a curvilinear
relationship between heterogeneity and per-formance.^42 An
example of such a recommendation is that “groups should pay
careful attention to how much they increase diversity... too
much diversity can lead to communication problems and
unavoidable conflict.”^43 Unfortunately, if one had knowledge of
an optimal level of heterogeneity, the application of such
information to increase performance might lead to specific
consideration of race, ethnicity, or gender in making work
assignments, which would appear to violate civil rights