Multiple hurdle selection strategies
Finch, David matthew
MetadataShow full item record
A Monte Carlo simulation is conducted to illustrate the tradeoffs in predictive efficiency and adverse impact associated with various predictor combinations in multiple hurdle selection strategies. Using meta-analytic estimates of intercorrelations, validity, and subgroup mean differences, twenty-eight multiple hurdle combinations of cognitive ability and three popular alternative predictors, namely biodata, conscientiousness, and the structured interview, are simulated across five selection ratios. Current results illustrate that multi-stage strategies are clearly more effective than single-stage strategies for balancing the predicted performance and minority hiring goals. Additionally, levels of adverse impact and mean performance will vary substantially depending upon the sequential ordering of predictors. This study increases our understanding of the tradeoff effects associated with multi-stage selection strategies. This increased understanding can enable practitioners to make better decisions regarding what type of strategy is best suited for their specific selection goal, particularly when a highly diverse and high performing workforce is the objective.