This research improved on previous studies by using demonstrated high-risk cases rather than projected risk, analyzing individual personality profiles rather than aggregate scale scores, and employing a broader dimensional conception of psychopathy. Participants were drawn from a multisite evaluation of batterer programs that involved 840 men admitted to 4 batterer programs (210 at each site). The programs were located in Pittsburgh, Houston, Dallas, and Denver. The demographics and behavioral characteristics of the full sample were similar to previous batterer samples used to derive typologies. Men were classified as "repeat reassaulters" if a partner reported physical violence against her more than once during a single 3-month follow-up. To minimize misclassification in reassault types, women's reports were verified against men's reports and police records. The MCMI-III (Millon Clinical Multiaxial Inventory, Version III) was used to assess personality and psychopathology among the subjects. The personality profiles of 580 men who did not reassault their partners, reassaulted once, or repeatedly reassaulted during a 15-month follow-up were grouped. Sixty percent of the "repeat reassaulters" showed no serious personality dysfunction or psychopathology. Only 11 percent had profiles suggesting conventional or primary psychopathic disorder. The broadest conception of psychopathy, including secondary psychopathy and both psychopathic disorder and style, applied to 54 percent of the repeaters. There was no significant difference among the reassault types for personality dysfunction, psychopathic disorder, or personality type; however, a significantly greater portion of repeaters showed some psychopathic tendency. These findings contradict the generalizations about high-risk batterers and imply that batterer counseling may be appropriate for many of them. 1 figure, 3 tables, and 40 references