Estimating Choice Probabilities in Multiattribute Decision Making

Abstract

The Bradley-Terry-Luce (BTL) model is adapted and applied to choice problems arising in multiattribute decision making, and is compared with three other models in an experiment dealing with employer choice in a job-seeking context. Results of the experiment indicate that the BTL adaptation provides a reasonably accurate model of subjects' subjective probabilities, as elicited by the constant-sum scaling technique.