Cognitive Differentiation Analysis: A Regression Extension of the Reynolds-Sutrick Model

Cognitive Differentiation Analysis (CDA) represents a method to directly measure the correspondence of an individual vector or a composite vector of descriptor ratings to a matrix of pairwise dissimilarity judgments where both sets of judgments are assumed to be ordinal. The zero intercept regression extension of CDA detailed in this paper yields a weight for each descriptor in the set, as opposed to the standard equal weighting case from which a composite vector would be constructed. An artifical example and an applied example are presented utilizing both ordinal and interval assumptions for the descriptor variables.