Page 51 - Psychological Management of Individual Performance

30 ability and non-ability predictors of job performance
TABLE 2.1 Meta-analytic results for cognitive ability–performance relations
Cognitive ability
Reference k N r

Dunnette (1972) a Performance — — .45
Pearlman et al. (1980) abc Proficiency 194 17539 .52
Training 65 32157 .71
Schmidt et al. (1980) bcd Proficiency 42 1299 .73
Training 9 1635 .91
Callender and Osburn (1981) cd Performance 37 3219 .32
Training 14 1694 .54
Schmidt et al. (1981) bd Performance 16 1486 .26
Hunter (1982) a Ratings 10 — .48
Vineberg and Joyner (1982) a Ratings 11 .28
Hunter and Hunter (1984) bcde Proficiency — — .45
Training — — .54
Schmitt et al. (1984) e Performance 53 40230 .25
Schmitt et al. (1997) Performance — — .29
Vinchur et al. (1998) cd Ratings 25 1770 .31
Bobko et al. (1999) Performance — 41553 .30
Colquitt et al. (in press) bce Skill acquisition 17 6713 .38
Notes: k = number of correlations; N = number of individuals across k samples; r = estimated true-score
relation.
a
Reanalyzed and reported in Hunter and Hunter (1984).
b
Estimated true-score relation corrected for predictor unreliability.
c
Estimated true-score relation corrected for criterion unreliability.
d
Estimated true-score relation corrected for range restriction.
e Estimated true-score relation corrected for sampling error.
Nonetheless, many I/O researchers continued to regard measures of general cognitive
ability as efficient predictors of job performance and occupational success. During the
late 1970s and 1980s, the introduction of the validity generalization approach to I/O
psychology research by Schmidt, Hunter and their colleagues (e.g., Hunter et al., 1982)
provided a method by which researchers could aggregate results across studies to esti-
mate true-score relations. These meta-analytic studies of ability–performance relations
provided empirical evidence indicating the predictive validity of general cognitive ability
measures for job performance across a wide range of jobs.
Our review of the literature yielded 13 meta-analytic studies on cognitive ability–
performance relations published over the past two decades. Table 2.1 presents a sum-
mary of the results of these meta-analytic studies on ability–performance relations, or-
ganized by study and further by criterion type (overall performance, supervisor ratings,
proficiency, training, skill acquisition). (Values reported in Table 2.1 represent esti-
mated true-score relations, with values corrected for predictor unreliability, criterion
1
unreliability, range restriction, and sampling error as noted). The magnitude of ob-
tained estimated true-score relations is hypothesized to vary by criterion type. That is,
while overall performance and supervisor ratings represents multidimensional outcomes
defined generally as observable behavior relevant to organizational goals (Campbell,
McCloy, Oppler, & Sager, 1995), job proficiency is generally more narrowly defined
as a task-relevant outcome (e.g., Schmidt, Gast-Rosenberg, & Hunter, 1980). Training