The authors propose a methodology for determining the segment-level impact of explanatory variables on multiple criterion measures obtained on a constant-sum scale. These explanatory variables could characterize different product, situation, or person related conditions that either occur naturally or are experimentally manipulated. Their proposed methodology simultaneously estimates market segment membership and multivariate segment-level parameters for each dependent criterion, using finite mixtures of conditional Dirichlet distributions. They conduct a modest Monte Carlo simulation analysis to investigate the performance of the proposed methodology. The authors also provide an empirical application to industrial buying decisions that examines the impact of the type of buying situation on multiple vendor selection criteria such as economic cost, functional performance, vendor cooperation, and vendor capability.
PCovR2: A flexible principal covariates regression approach to parsimoniously handle multiple criterion variables
