Page 66 - Psychological Management of Individual Performance

progress, problems, and prospects for future research 45
In a related vein, apparent inconsistencies in the accumulated meta-analytic findings
have not been fully explored. Ackerman (2000), for example, found that individual dif-
ferences in extroversion were negatively related to measures of fluid and crystallized
intelligence. However, findings by several researchers (e.g., Barrick & Mount, 1991)
indicate a positive relationship between extroversion and job performance. If general
cognitive ability is positively related to job knowledge (as suggested by Schmidt et al.,
1986) and job performance, and extroversion is positively related to job performance,
then what is the relationship between extroversion and job knowledge? If a negative
relationship between extroversion and job knowledge is obtained, there may be an in-
teraction or asymmetric trade-off function between intellect and sociability factors in
the prediction of job performance. Persons high in both general cognitive ability and
extroversion, for example, might develop less job knowledge compared to persons high
in general cognitive ability but low in extroversion. To the extent that the development of
job knowledge contributes significantly to job performance, such as in training contexts,
higher levels of extroversion may dampen the positive influence of general ability on
performance.

PROSPECTS

In their review of recent person predictor research, Mount and Barrick (1998) note:
“... there are now two dispositional predictors in our field whose validity generalizes:
general mental ability and conscientiousness. Thus, no matter what job you are selecting
for, if you want employees who will turn out to be good performers, you should hire
those who work smarter and work harder” (p. 856). We concur that over the past two
decades, I/O psychologists have provided critical empirical evidence to support long-
standing assumptions about the broad validity of general mental ability and non-ability
traits for prediction of job performance. As such, it can be argued that with these issues
largely settled, contemporary research efforts need only tie up loose ends.
We disagree, and suggest that such a conclusion represents what the physicist Feynman
(1985) describes as “cargo cult science”, a term based on Feynman’s observation of the
failure of a group of island natives in the Pacific Ocean to reinstate cargo shipments of
supplies to their island after the end of WWII despite their extensive efforts to refurbish
and staff the abandoned landing strip. As Feynman (1985) noted, the islanders simply did
not recognize that the reason the cargo planes did not return to the island was a function
of the war’s end, and had nothing to do with the precision with which the islanders
recreated the landing environment.
As described by Feynman (1985), cargo cult science is characterized by slavish ad-
herence to scientific procedures in the absence of theoretical understanding and the
development of alternative explanations for observed effects. According to Feynman
(1985), such approaches ultimately fail in terms of advancing scientific knowledge or
understanding. Meta-analytic findings over the past 15 years have produced provocative
results. But to avoid the consequences of cargo cult science, we suggest that future re-
search will need to focus greater attention on answering fundamental questions about
the relationships among predictor and criterion constructs, and providing alternative
explanations for variability in predictor–performance relations.