A Monte Carlo Study of Metric and Nonmetric Estimation Methods for Multiattribute Models

Several procedures for estimating the parameters in a model of preference formation have been developed and applied in the past decade. Four methods are compared based on synthetic rank order data. The results indicate that one nonmetric procedure suffers greatly from “local optimum” problems, and is only marginally different in performance from a well-known metric procedure. Conditions are identified under which another nonmetric method outperforms metric procedures.