Abstract

This study evaluated the extent to which the features of personality disorders are organized into the 11 diagnoses proposed by DSM-III-R. The traits composing personality disorder diagnoses were identified in earlier studies. Seventy-nine traits were required to represent personality diagnoses. Self-report scales were developed to measure each trait. The factorial structure underlying the trait scales was examined in a sample of 158 patients with a primary diagnosis of personality disorder and a sample of 274 general population subjects. Eleven principal components were extracted from each data set and were rotated to the same oblique criterion. The decision to extract 11 components was based on a priori expectations derived from DSM-III-R. Factor comparability coefficients were computed to quantify the similarity of the 2 solutions. A high degree of similarity was observed between the factor structure in the clinical and general population samples, suggesting that features of personality pathology are continuous dimensions. Although a number of factors demonstrated a degree of resemblance to some DSM-III-R diagnostic categories, the overall correspondence was not strong. A second-order factor analysis failed to reproduce the 3 Axis II diagnostic clusters. The results do not support the categorical model used in DSM-III-R and they provide only limited support for DSM-III-R diagnostic concepts.