Cognitive Ability Tests

Prediction of Workplace Outcomes

Cognitive ability tests have long been used in industry to assess preemployment candidates because of the strength with which they predict on-the-job performance for all jobs. In fact, extensive meta-analytic evidence from validity studies across a range of jobs (e.g., clerical, military, sales, and white-collar positions) makes a strong case for g, or general mental ability (GMA), as the single best predictor of performance. Meta-analytic studies conducted in the United States report predictive validity between GMA and performance ranging from .31 (refinery workers) to .73 (computer programmers), and meta-analytic studies from the European Community report a similar range and an overall operational validity of .62.

As evidenced by this wide range of correlations in the United States and in Europe, the value of the predictive validity of GMA for job performance varies according to the complexity of the job. For example, manual labor jobs tend to have validity coefficients around .25, whereas most white-collar jobs tend to have coefficients ranging from .50 to .60. Therefore, it is the complexity of the job that moderates the validity of GMA for predicting job performance. Along with job complexity, another factor originally thought to be related to the GMA-performance relationship is job experience. In particular, it was thought that this relationship would decrease with increasing job experience. Recent investigation, however, shows that the predictive validity of GMA is stable and does not decrease over time with increased job knowledge. This finding makes sense because GMA is believed to be the factor that turns experience into increased performance.

In addition to their use in predicting job performance, cognitive ability tests have also been used to successfully predict other important workplace criteria. For example, U.S. meta-analytic research findings show a strong predictive correlation of .62 between [Page 80]GMA and training performance, whereas European findings report a correlation of .54. Meta-analytic research also shows that GMA is related to occupational attainment, though findings vary depending on whether the investigation is cross-sectional (.62) or longitudinal (.51). Cognitive ability results increase and standard deviations and score ranges decrease with increasing occupational level; in other words, it seems that those low in GMA have a more difficult time attaining high-level occupations.

...