Empirical clusters of DSM-III personality disorders in violent offenders

J Pers Disord. Spring 1999;13(1):18-34. doi: 10.1521/pedi.1999.13.1.18.


Violent male offenders in a maximum security hospital and special units in prisons (N = 164) were interviewed with the Structured Clinical Interview for DSM-III Axis II disorders (SCID-II). Cluster analysis of the personality disorder criteria sets identified six diagnostic patterns: (1) antisocial-narcissistic; (2) paranoid-antisocial; (3) borderline-antisocial-passive-aggressive; (4) borderline; (5) compulsive-borderline; and (6) schizoid. Offenders in the first three groups had more extensive criminal careers, and most were identified as psychopaths by the Psychopathy Checklist-Revised (PCL-R). These Groups also had more frequent lifetime histories of substance abuse. A history of affective and anxiety disorders was more common in Groups 3 and 5, and almost two thirds of Group 2 had a history of psychotic disorder. The results emphasize that dangerous offenders are heterogeneous in personality pathology. They also suggest that personality disorder among violent offenders is more commonly represented by recurring patterns of covarying traits than by single categorical entities proposed in the DSM classification.

MeSH terms

  • Adult
  • Antisocial Personality Disorder / classification
  • Antisocial Personality Disorder / diagnosis
  • Antisocial Personality Disorder / epidemiology
  • Cluster Analysis
  • Comorbidity
  • Crime / statistics & numerical data
  • Hospitalization
  • Humans
  • Male
  • Mental Disorders / diagnosis
  • Mental Disorders / epidemiology
  • Personality Disorders / classification
  • Personality Disorders / diagnosis
  • Personality Disorders / epidemiology*
  • Personality Inventory / statistics & numerical data
  • Prisoners / psychology
  • Prisoners / statistics & numerical data*
  • Psychiatric Status Rating Scales / statistics & numerical data*
  • Psychotic Disorders / diagnosis
  • Psychotic Disorders / epidemiology
  • Terminology as Topic
  • Violence / psychology
  • Violence / statistics & numerical data*