Latent Class Multidimensional Scaling. A Review of Recent Developments in the Marketing and Psychometric Literature

We present a description of current, state-of-the-art methods for per­forming latent class multidimensional scaling (LCMDS). The organizing framework utilized to present these various methods revolves around the types of data collected and the associated spatial representation desired from such approaches. As such, methods for proximity, paired compar­isons, preference, choice, and profile or ratings data are described. A general form of the E-M maximum likelihood estimation procedure typ­ically utilized for the vast majority of these methods is also presented. Finally, the potential application of these methods in marketing, as well as directions for future research, are discussed.