L. Ekselius et al., A PRINCIPAL COMPONENT ANALYSIS OF THE DSM-III-R AXIS-II PERSONALITY-DISORDERS, Journal of personality disorders, 8(2), 1994, pp. 140-148
By means of the SCID Screen questionnaire, data concerning 76 reported
criteria were collected from 176 controls and 388 patients. In a prin
cipal component analysis, 23 factors emerged that explained 62.7% of t
he variance. After regrouping, a pattern similar to the Clusters A, B,
and C could be demonstrated. However, the criteria for Borderline Per
sonality Disorder scattered in all three clusters. After rearrangement
according to the loadings in components 1 and 2, two clear axes could
be demonstrated. One axis went from overconscientiousness and inflexi
bility in one extreme, to unstable intense relationships in the other.
The second axis went from avoiding social contacts in one extreme end
, to uncomfortable if not in the center of attention in the other. Whe
n the separate personality disorders were arranged according to factor
loadings in component 1 and 2, Borderline and Narcissistic Personalit
y Disorders grouped together whereas Histrionic Personality Disorder f
ormed a separate group. Schizotypal, Schizoid, and Paranoid Personalit
y Disorders grouped together and so did Avoidant, Obsessive-Compulsive
, and Dependent Personality Disorders; Passive-Aggressive Personality
Disorder grouped together with the Cluster A disorders.