Latent structure of DSM-III-R Axis II psychopathology in a normal sample.

The Personality Disorder Examination was administered to 302 normal controls in the New York High-Risk Project in order to elicit Axis II diagnoses (revised 3rd edition of the Diagnostic and Statistical Manual of Mental Disorders; American Psychiatric Association, 1987) and quantitative dimensions of psychopathology. LISREL confirmatory factor analysis was used to evaluate the Axis II hypothesis of 3 orthogonal factors. There was considerable overlap among personality disorders. The best fitting LISREL model was of 3 oblique factors that were different for male and female subjects. Given that our choice of variables to constrain in order to mathematically identify our models was partially based on analysis of intercorrelations in our data set, our methods were not purely confirmatory. We present our results not to confirm specific hypotheses but to generate explicit hypotheses that can be tested in independent samples.