The use of smallest space analysis in studying scale structure: An application to the California Psychological Inventory.

Smallest space analysis is a nonmetric technique for analysis of proximity relations, e.g., intercorrelation matrices. The variables are represented as points in Euclidean space so that the rank order of the interpoint distances corresponds to the rank order of the intercorrelations. A nonmetric technique developed by L. Guttman and J. C. Lingoes, G-L SSA-1, was applied for analysis of the structure of the intercorrelation matrix of the CPI. 2 dimensions provided a good fit to the data. The space can be partitioned into several regions. The central region includes scales related to general adjustment. 3 peripheral regions include measures of "person orientation," "value orientation," and "self-orientation." Results correspond to those obtained in factor analytic studies, but the nonmetric solution is simpler and more parsimonious. Applications to prediction problems, scale construction, and formulation of research hypotheses are discussed. (18 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved)