Simultaneous parameter estimation for the multiplicative multiattribute utility model

Abstract

A nonlinear least squares procedure for simultaneous estimation of the parameters of the multiplicative utility functions (R. L. Kenney & H. Raiffa, Decisions with Multiple Objectives: Preferences and Value Trade-offs, New York: Wiley, 1976) is developed. The procedure, SEE, improves on Barron and Person, 1979, Keeney, 1974 in that it allows one to treat the measurement error consistently and to test a variety of hypotheses by standard methods of statistical inference. Numerical examples are studied to test the performance of SEE. It is concluded that SEE is feasible and that it leads to correct conclusions about a number of interesting hypotheses, when certain conditions on the error term are satisfied.