Optimal Designs for Stated Choice Experiments Generated From Fractional Factorial Designs

Abstract

This article gives optimal designs obtained by developing a fractional factorial design for the estimation of main effects in stated choice experiments under the assumption of equal selection probabilities. This construction approach follows that of Burgess and Street (2005), who develop complete factorial designs to construct optimal designs for choice experiments, but we obtain choice experiments with fewer choice sets. We construct the fractional factorial designs using the Rao-Hamming method, which assumes all attributes have the same number of levels, which must be a prime or a prime power. We also find optimal designs for stated choice experiments that are generated from asymmetric fractional factorial designs constructed using expansive replacement under the same assumption. We use the multinomial logit model to analyze the results, and we make the assumption of equal selection probabilities when calculating optimality properties. The methods that we use to implement these constructions are given in the last section.