The initial development is described of a quantitative system that provides for differential representations of perceptual structures for different individuals. A system is desired which would provide not only descriptions of individual perceptual structures and a basis for comparisons between individuals and groups, but also a superstructure with which to gain an understanding of the variety of individual perceptual structures. Richardson's multidimensional scaling model is used, with the Young and Householder theorem, and extensions by Torgerson and Messick. Measures of dissimilarity for pairs of stimuli are entered in a matrix to which a type of factor analysis is applied; all cells in the matrix must be filled. The analysis next uses Eckart and Young's procedure; the analysis is applied to the matrix of the sums of squares and products of raw measures. The results involve a row of matrix for each real individual and another row of matrix for each pair of stimuli. In a study of political thinking, students were divided into four groups representing liberals or conservatives in favor of either labor or management. Three dimensions were used in the structures of individual differences. A shift was noted in the complexity of perceptual structures from simple spaces for individuals toward a complex space for an idealized individual.