Page 299 - Discrimination at Work The Psychological and Organizational Bases

STONE-ROMERO
266
CONSTRUCT VALIDITY PROBLEMS IN MEASURES
OF NEGATIVE AFFECTIVITY
The use of measures of NA for various organizational purposes would
not be problematic if such measures had substantial levels of construct
validity. However, as is detailed below, there are serious construct validity
problems with such measures.
Uses of the Negative Affectivity Construct in Organizational Research
Taken at face value, the arguments in support of the trait nature of NA
seem compelling. So much so, it appears, that entire programs of research
in industrial and organizational psychology and related fields have been
devoted to demonstrating the value of NA for predicting a host of variables
that are of relevance to work organizations (e.g., Brief et al., 1988; Burke,
Brief, & George, 1993; Levin & Stokes, 1989). In addition, as noted above,
two major meta-analytic reviews have considered the criterion-related va­
lidity of emotional stability (equivalent to NA) for predicting such crite­
ria as job performance, tenure, promotions and other status changes, and
salary (Barrick & Mount, 1991; Tett et al., 1991). As a consequence of re­
search on NA, several publications have presented overly optimistic and
highly misleading views of its value in predicting and/or explaining or­
ganizationally relevant variables. For example, George and Jones (2004)
wrote that "Individuals who are high on neuroticism [or NA] are more
likely to experience negative moods at work, feel stressed, and have a
negative orientation to the work situations" (p. 9). In addition, DuBrin
(2004) argued that "People with negative affectivity are often distressed
even when working under conditions that coworkers perceive as inter­
esting and challenging" (p. 164). Moreover, Cascio (1998) stated that "the
evidence now indicates that scores on well-developed measures of normal
personality (1) are stable over reasonably long periods of time, (2) predict
important occupational outcomes,... [and] ... (3) do not discriminate un­
fairly against any ethnic or national group" (p. 228). As is demonstrated
below, there are very serious construct validity problems with many of the
studies that have dealt with relations between NA and various outcomes.
Thus, the validity of the just-noted arguments (Cascio, 1998; DuBrin, 2004;
George & Jones, 2004) is highly debatable.

Confounding of Measures of Negative Affectivity

A review of many popular measures of NA (and of such related constructs
as neuroticism, anxiety, and depression) reveals a very serious construct