Utility Analysis for Multiple Selection Devices and Multiple Outcomes

Abstract

Traditional utility analysis only calculates the value of a given selection procedure over random selection. This assumption is not only an inaccurate representation of staffing policy but leads to overestimates of a device's value. This paper generates a new utility model that accounts for multiple selection devices and multiple criteria. The model is illustrated using previous utility analysis work and an actual case of secretarial employees with eight predictors and nine criteria. A final example also is provided which includes these advancements as well as other researchers' advances in a combined utility model. Results reveal that accounting for multiple criteria and outcomes dramatically reduces the utility estimates of implementing new selection devices.