Page 305 - Discrimination at Work The Psychological and Organizational Bases

STONE-ROMERO
272
The Correspondence Bias in Person Perception
Research in social cognition shows evidence of a correspondence bias in per­
ceptions of targets. More specifically, on the basis of their observations of
targets, observers will infer that targets have traits or dispositions, even
when the behavior of the targets was a function of or was constrained by
situational factors (Hamilton & Sherman, 1994). For example, observers are
likely to attribute signs of stress or strain that are observed in targets to un­
derlying dispositions, as opposed to the effects of environmental stressors.
To the degree that dispositional inferences are made about targets, they are
likely to suffer both access and treatment discrimination. Thus, not only
will targets suffer harm from a host of other environmental factors (e.g.,
stressful jobs, poor housing), but also they will further be harmed by the
correspondence biases of observers.

Unfair Discrimination on the Basis of Strain-Confounded Measures

Because of the above-noted relations among race, SES, exposure to stres­
sors, and strain, to the extent that measures of psychological strain (e.g.,
neuroticism, NA, emotional stability) are used in personnel selection, in­
dividuals who have elevated scores on such measures may suffer both
access and treatment discrimination. More specifically, the use of such mea­
sures for selection purposes may lead to indirect race-based discrimination
against members of several minority groups (eg., African Americans, Mex­
ican Americans, American Indians). In addition, if such measures are ap­
plied to existing employees, those who have elevated scores on them may
suffer various types of treatment discrimination. Overall, therefore, it ap­
pears highly inappropriate to use measures of psychological strain (e.g.,
many popular measures of NA) for most personnel decision-making pur­
poses. However, there is one important exception to this. More specifically,
it is acceptable to use information from such measures for making decisions
about providing individuals with access to employee assistance programs
(e.g., stress reduction programs).
Although there is considerable evidence of relations among SES, race
and psychological strain, some research suggests that the use of measures
of personality for selection purposes generally does not result in adverse
impact. More specifically, based upon the analysis of data derived from em­
pirical studies of several other researchers, Hough, Oswald, and Ployhart
(2001) reported that there were very small difference between Blacks and
Whites on measures of adjustment (standardized mean score difference, d
less than .05). Interestingly, however, relative to Whites (a) Blacks had lower
scores on measures of affiliation, dependability, openness to experience,