Determining Predictor Weights in Military Selection: An Application of Dominance Analysis |
| |
Authors: | Wendy A Darr Victor M Catano |
| |
Institution: | 1. Directorate of Military Personnel Research and Analysis, Defence Research and Development Canada, Ontario, Canadawendy.darr@forces.gc.ca;3. Department of Psychology, Saint Mary’s University, Nova Scotia, Canada |
| |
Abstract: | This article illustrates the use of dominance analysis (DA) to identify the most appropriate set of weights for combining predictor data. We used meta-analytic information on cognitive ability, structured interviews, conscientiousness, and emotional stability to determine relative weights in predicting entry-level job performance and turnover. For job performance, the optimal predictor weights were 51% (cognitive ability), 38% (structured interview), 9% conscientiousness, and 2% (emotional stability). The weights differed considerably for turnover: 48% (structured interview), 27% (emotional stability), 13% (conscientiousness), and 11% (cognitive ability). Using a simulated data set, we showed how the rank-order of applicants can change based on DA and traditional regression analysis. The results suggest that DA has wide applicability for military selection. |
| |
Keywords: | differential weights selection composite meta-analysis dominance analysis |
|
|