Within the framework of Holland's theory, we examined the use of person-env
ironment congruence in predicting job performance for a sample of customer
service representatives. it was predicted that: (1) Congruence scores based
on specific environment classification derived by job analytic methods wou
ld correlate more highly with performance than congruence scores based on a
more general environmental classification; (2) Congruence and cognitive ab
ility would correlate with different aspects of performance (task v. contex
tual performance); and (3) Congruence scores would significantly relate to
job performance, whereas interest test scores would not. No support for the
first hypothesis was found. Instead, there were no significant differences
between the environment typing methods, and congruence predicted Quality p
erformance ratings. Partial support was found for the second and third hypo
theses. Congruence was significantly correlated with task-related ratings,
whereas cognitive ability was not significantly correlated with any perform
ance ratings. In addition, none of the three relevant interest test scores
(Conventional, Social, and Enterprising) were significantly correlated with
performance. A significant negative correlation was found, however, betwee
n performance ratings and investigative interest scores, suggesting a need
for person-environment fit indices to rake into account all six interest sc
ores. Overall, this study provides some evidence that person-environment fi
t, when guided by theory, may be a useful predictor of job performance. (C)
1999 Academic Press.