Weighted composites of personality facets: An examination of unit, rational, and mechanical weights

Past research indicates that personality facets within the same factor are often differentially related to criteria. This implies that differential weighting of facets may be advantageous. We considered unit weighting, rational weights, and mechanical weights in terms of empirical prediction and theoretical understanding. Two studies indicate that regression weighted personality facet composites performed best. Rational weights performed moderately well and unit weights performed worst. These findings suggest that regression weights are best when sample sizes are large, and rational weighting offers moderate prediction without the need for a large empirical study. We recommend using facet-level personality measures to take advantage of the predictiveness and theory development that can be gained from studying facet-criterion weights and relations.