Multidimensional Scaling: An Introduction and Comparison of Nonmetric Unfolding Techniques

Abstract

First this article describes some of the basic concepts of multidimensional scaling of similarities and preference data and provides a short description of its historical development. Then it reports an empirical comparison of three computer programs for unfolding preference data to “recover” a stimulus configuration (independently obtained). Results indicate differences among the resulting configurations that reflect the influence of differential weighting of “perceptual” dimensions in the context of preference.