Fairness In Personnel-Selection - A Metaanalysis And Policy Implications

Previous research has suggested that selection procedures and performance evaluations are not unfair to minorities. Results of a meta-analysis indicated that Whites performed higher than non-Whites on cognitive ability tests (dBAR = .464) and on supervisory ratings (dBAR = .284), but not on objective results (dBAR = -.009) and that validities between the tests and ratings and results were not significantly different. In addition, a comparison of predicted to actual mean standardized criterion differences between White and non-White subgroups suggested that cognitive ability underpredicted actual differences in supervisory ratings. These results imply that researchers and policymakers need to continue to evaluate fairness in testing and performance evaluation because, contrary to prior evidence, the use of cognitive ability tests may indeed be unfair to non-Whites.