A model-based approach for visualizing the dimensional structure of ordered successive categories preference data

A cyclical conditional maximum likelihood estimation procedure is developed for the multidimensional unfolding of two- or three-way dominance data (e.g., preference, choice, consideration) measured on ordered successive category rating scales. The technical description of the proposed model and estimation procedure are discussed, as well as the rather unique joint spaces derived. We then conduct a modest Monte Carlo simulation to demonstrate the parameter recovery of the proposed methodology, as well as investigate the performance of various information heuristics for dimension selection. A consumer psychology application is provided where the spatial results of the proposed model are compared to solutions derived from various traditional multidimensional unfolding procedures. This application deals with consumers intending to buy new luxury sport-utility vehicles (SUVs). Finally, directions for future research are discussed.