Further Consideration of the Use of Narrow Trait Scales

This study examined whether inclusion of narrow personality trait scales improved prediction of job performance enough to overcome increased capitalization on chance characteristics of the sample. Data were obtained from 262 midlevel managers who had completed a personality inventory and had been assessed using multisource performance ratings. Three methods of combining trait scores organized under the 5-factor model were compared in order to better understand the contribution of variance specific to the narrow traits in the prediction of performance: common factor scores, unit-weighted composite scores, and regression-weighted composites. Results showed that on average the sets of facet scores explained an additional 9% of the variance in performance beyond summated composites and 10% beyond the factor scores (4% and 5% after correction for shrinkage). Contrary to the idea that broadband predictors would fare better with broad criterion constructs, the increment was similar whether the criterion was a single dimension of performance or overall job performance. Recommendations include reporting results from multiple levels of the personality hierarchy, and suggestions are made for situations where sample size limits the number of predictors advisably entered into a regression equation.