Page 289 - Discrimination at Work The Psychological and Organizational Bases
P. 289
STONE-ROMERO
256
(d) the consequences of stigmatization, (e) the personality construct, (f) per
sonality measures in organizational contexts, (g) the negative affectivity
construct, (h) construct validity problems with measures of negative af
fectivity, (i) social causation models of well-being, (j) the criterion-related
validity of personality measures, and (k) several conclusions concerning
the use of personality measures as predictors of various criteria.
UNFAIR DISCRIMINATION IN THE WORKPLACE
Measures that discriminate have the capacity to differentiate between the
attributes of measured entities (e.g., people). The better the measure, the
greater its capacity to discriminate and provide information about the de
gree to which measured entities differ from one another. For example, a
meter stick that is calibrated in millimeter units has a greater capacity to
discriminate among the heights of individuals than one that is calibrated
in centimeter units. In addition, all else constant, a test of job-related ability
with 50 items has greater capacity to discriminate among test takers than a
test with 10 items. The longer test would have greater reliability and, thus,
greater ability to discriminate.
Fair Discrimination
There is nothing wrong (vis-a-vis either psychometric or ethical perspec
tives) with measures that are valid and, thus, discriminate fairly. For ex
ample, there would be no problem with an organization using a highly
discriminating, construct valid (i.e., highly reliable and unbiased) mea
sure of job aptitude for the purpose of personnel selection. In fact, all else
constant, the greater the capacity of such a measure to discriminate, the
more effective it would be in predicting job performance. However, a bi
ased measure would present both operational and ethical problems. More
specifically, because of the fact that it lacked construct validity, it would
discriminate unfairly.
Unfair Discrimination
In organizational contexts, unfair discrimination can result from the use of
invalid (biased) measures of both predictors and criteria. For example,
a preemployment test that measured factors that were not predictive of
actual job performance (e.g., previous membership in college fraternities
or sororities) would discriminate unfairly among job applicants. In addi
tion, a measure of job performance would discriminate unfairly against job