MathSoft, Inc. - Data Analysis Products Division
Recent algorithms for nonlinear mixed effects models can also be used, after minor modification, to fit individual differences multidimensional scaling models in which the subject weights are random effects. These models, which we propose here for the first time, have several advantages over models which do not use random effects. For example, unlike traditional models, the number of parameters does not increase with the number of subjects, and, because the distribution of the subject weights is modeled, generalization to the sampled population of subjects is immediate. These models also have several disadvantages. Here we discuss the proposed random effects model, give a computational algorithm for fitting these models, describe our experiences with this algorithm, and discuss potential generalizations. The talk will begin with a tutorial on multidimensional scaling.