Multidimensional scaling (MDS) is a psychometric method with wide application in behavioral science research. We propose to develop software for a new class of MDS models. In these new models parameters associated with individuals are modeled as random effects rather than as fixed parameters. For the diagonal metric (or INDSCAL) models, these parameters are the subject weights. The resulting random effects MDS model has many advantages over its classical counterpart. For example, we are better able to estimate subject weights even when only one dissimilarity is observed on an individual, and we can make model-based inferences about the sampled population of subject weights. We propose to develop a comprehensive module of computational algorithms for computing estimates in this new class of MDS models. Included in this module will be software for model fitting, inference, diagnostics, and other appropriate statistical techniques, a graphical user interface, a users manual, and online documentation. The software will also contain procedures for robust estimation.
Multidimensional scaling is a psychometric method widely used in market research, psychology, sociology, political science, genetics, chemometrics, and other areas of behavioral and scientific research. The proposed models have significant advantages over existing techniques. A well designed and comprehensive module for implementing these models will find a ready market.
