Narrow reasoning about the use of broad personality measures for personnel selection

Ones and Viswesvaran (1996) have argued for the use of broad bandwidth, rather than narrow bandwidth, personality predictors in personnel selection research when overall job performance is the criterion of interest. We take the opposite position in this article—that homogeneous measures of unidimensional personality traits are always to be preferred as predictors of work (and other) criteria. We maintain that the use of multiple unidimensional predictors provides important advantages over the use of multidimensional aggregates of those predictors. These advantages pertain to both (a) empirical accuracy in predicting job performance, and (b) psychological meaningfulness in explaining work behavior. Our conclusions are supported by Ones and Viswesvaran's own data.