THE STUDY GATHERED MINNESOTA MULTIPHASIC PERSONALITY INVENTORY PROFILES (MMPI) AND DEMOGRAPHIC DATA ON 6 GROUPS OF 25 MEN EACH WHO HAD BEEN REMANDED BY THE COURTS TO A MAXIMUM SECURITY PSYCHIATRIC HOSPITAL FOR PRETRIAL ASSESSMENT. IN ADDITION TO THE MMPI SCORES, THE FOLLOWING DATA WERE COLLECTED: NUMBER OF GRADES COMPLETED, AGE ON ADMISSION TO A MENTAL HOSPITAL OR CORRECTIONAL INSTITUTION PRIOR TO THE CURRENT OFFENSE, MARITAL STATUS, ADMISSION DIAGNOSIS, VOCABULARY SUBSCALE, AND PERFORMANCE IQ. THE GROUPS DIFFERED ACCORDING TO OFFENSE TYPE: MURDER OR ATTEMPTED MURDER OF A FAMILY MEMBER OR GIRLFRIEND, MURDER OR ATTEMPTED MURDER OF AN UNRELATED VICTIM, RAPE, ARSON, CHILD MOLESTING, OR PROPERTY OFFENSE. THE MMPI PROFILES OF THE GROUPS WERE REMARKABLY SIMILAR, AND THE PREDICTION THAT THE MURDER FAMILY GROUP WOULD HAVE THE HIGHEST OVERCONTROLLED HOSTILITY (O-H) SCORES WAS NOT SUPPORTED. MULTIPLE DISCRIMINANT ANALYSES SHOWED THAT THE REMAND'S AGE ON ADMISSION AND HIS EXPERIENCE IN CORRECTIONS BEFORE THE CURRENT OFFENSE WERE THE BEST DISCRIMINATORS AMONG THE GROUPS. THESE DATA INDICATE THAT THE MMPI IS NOT RELATED TO THE OFFENSE TYPES OF MENTALLY DISORDERED OFFENDERS REFERRED BY THE COURTS FOR ASSESSMENT, BUT IT DOES REFLECT THEIR RELATIVELY HIGH LEVELS OF PSYCHOPATHOLOGY. IT IS SUGGESTED THAT O-H SCORES SHOULD BE USED WITH VERY GREAT CAUTION IN ASSESSING THE DANGEROUSNESS OF OFFENDERS. TABLES OF SCORES, DISCRIMINANT ANALYSES, AND CORRELATIONS ARE INCLUDED. APPROXIMATELY 20 REFERENCES ARE GIVEN. (AUTHOR ABSTRACT MODIFIED--MFE)