Page 343 - Discrimination at Work The Psychological and Organizational Bases

ARTHUR AND DOVERSPIKE
310
an HR perspective, the process of selection begins once the applicant is suc­
cessfully attracted to the organization, that is, they submit an application
or otherwise formally indicate an interest or desire to seek employment
with the organization. Next, some evaluative tool, usually a test (defined
here to include any assessment tool used in selection-related decision mak­
ing) is used to assist in selecting individuals from the pool of applicants.
In the testing and personnel selection literature, cognitively loaded paper-
and-pencil tests of knowledge, skill, and ability have been shown to be the
most valid predictors of job performance (Schmidt & Hunter, 1998). How­
ever, it has also been extensively documented that paper-and-pencil tests of
cognitive ability generally display large racial subgroup differences with a
widely cited one standard deviation difference in African American/White
performance. What remains unclear is whether the observed differences
are due to the construct being assessed (cognitive ability) or the method
of testing (multiple-choice paper-and-pencil tests; Arthur, Day, McNelly, &
Edens, 2003; Arthur, Edwards, & Barrett, 2002; Hough, Oswald & Ployhart,
2001; Schmitt, Clause, & Pulakos, 1996).
In addition, some aptitude tests, such as those of mechanical aptitude,
may result in substantial adverse impact as a function of sex. Paper-and-
pencil tests may also result in adverse impact toward older test takers.
The end result is that cognitively loaded paper-and-pencil tests of knowl­
edge, skill, ability, and aptitude, supposedly neutral devices, may in fact
represent a substantial barrier to the adequate representation of minorities
in organizations. This has resulted in a search for ways of reducing the
adverse impact associated with many traditional tests used in high-stakes
testing.
Since the ban on subgroup norming and other adjustments to test scores
on the basis of protected class status (Civil Rights Act, 1991), attempts to
reduce adverse impact have focused on a number of approaches includ­
ing (a) identifying and removing internal bias; (b) increasing test taking
motivation; (c) altering the selection criteria (weighting of tests, random
selection, race-based selection, banding); (d) changing the construct (the
use of nonability-based constructs including personality variables, infor­
mation processing skills and abilities, emotional intelligence, tacit knowl­
edge); and (e) changing the method or using alternative test formats (in an
attempt to alter test perceptions and attitudes and reduce nonjob-related
reading demands).
Identifying and Removing Internal Bias Identifying and removing inter­
nal bias from selection tests is predicated on the frequent critique that
ability tests are culturally biased. For about the past 40 years, psycholo­
gists have been trying to identify and eliminate biased items in an attempt
to improve tests. The techniques employed were at one time referred to as