Empirical reviews of personality measures conducted over the last 20 years have largely dispelled the myth that personality measures lack sufficient validity to be considered for use in applied settings (see Morgeson, Campion, Dipboye, Hollenbeck, Murphy, & Schmitt, 2007, for a uniquely dissenting opinion). Unfortunately, many personality job performance studies have adopted an atheoretical approach; researchers interpreted results without considering whether or not a given trait should be related to a targeted criterion, nor the direction of relationship (i.e., positive or negative). Thus, understanding the personality-performance relationship has been limited by the lack of confirmatory research strategies (Hogan & Holland, 2003; Schneider, Hough, & Dunnette, 1996; Tett et al., 1991, 1994, 1999) and an over-reliance on easily obtained criteria to represent job performance.;Two directions of interest in this study were (1) the value of confirmatory studies to enhance both the magnitude of prediction (Tett, Jackson, Rothstein, & Reddon, 1999) and the interpretability of results, and (2) the refinement of personality measures to accommodate the multidimensional nature of job performance. Specifically, this study attempted to enhance understanding of personality-job performance relations by using job performance taxonomies to create construct-aligned compound personality scales. These scales were constructed by mapping personality subdimensions to conceptually-related performance dimensions. Based on this strategy, the compound scales were expected to improve both the interpretability and strength of relationships between the two domains. The effectiveness of the compound scales was evaluated through meta-analysis and compared to meta-analytic results conducted on Big Five-based scales from the same studies.;Overall, with some qualifications, the study's hypotheses were supported. The average correlation across all criteria was twice as strong for the compound scales versus the average correlation for the Big Five-based scales for observed (.13 versus .06), operational (.17 versus .08), and true validity (.19 versus .09). These results build on previous research (Bartrum, 2005; Hogan & Holland, 2003; Tett et al., 1999) demonstrating improvements in predictive accuracy when conceptual linkages are established between personality traits and job performance by using specificity on both sides of the predictor-criterion equation to aggregate facets into compound scales.
展开▼
