This study used a policy-capturing approach to examine 3120 pay-alloca
tion decisions of 65 managers (31 male, 34 female) who work in organiz
ations that have a pay for performance policy. Subjects were asked to
respond to packages that varied in terms of performance of subordinate
s, specialized skills/expertise of subordinates, difficulty in replaci
ng the subordinates, possibility of potential subordinate turnover, an
d current salary of subordinates. Results indicated that all five fact
ors were significant and in the direction expected. The interaction be
tween performance and job offer was significant for only two of the 65
managers. The interactions between specialized skills and job offer a
nd difficulty in replacement and job offer were not significant for an
y of the managers. Most of the variance across the managers was accoun
ted by sampling error and criterion unreliability. Results also indica
ted that subjects' explicit rankings of the five factors did not corre
spond highly with the rankings of their regression weights.