Title
Conjoint Analysis for Volumetric Data
Speaker
Greg Allenby, The Ohio State University
Abstract
Conjoint analysis for stated preference data is developed within the context of constrained utility maximization using a model structure that nests the common discrete choice model. The model allows for multiple discreteness, characterized by a mixture of corner and interior solutions, and is extended for allocation data commonly encountered in the conjoint analysis of pharmaceuticals. We compare the proposed model to the standard logit model and regression model using data from a conjoint survey of physicians, and find that it results in consistently better predictive fits, particularly for predictions near zero and one. Application to conjoint data from a volumetric study of ice cream demand confirms the importance of the likelihood representing constrained utility maximization. Extensions to non-conjoint applications are discussed.
