Neuropsychological functioning was assessed in 39 males who had committed domestic violence (batterers) and compared to 63 nonviolent (both maritally discordant and satisfied) subjects recruited by advertisement. Subjects were subsequently divided into two groups (head injured, nonhead injured) and these groups were also contrasted as a function of batterer status. Tests were administered to assess for cognitive and behavioral functions, including executive dysfunction, hypothesized to be a factor contributing to propensity for violence. Questionnaires and structured clinical interviews were used to assess marital discord, emotional distress, and violent behaviors. Batterers differed from nonbatterers across several cognitive domains: executive, learning, memory, and verbal functioning. Batterers were reliably discriminated from nonbatterers based on three neuropsychological tasks: Digit Symbol, Recognition Memory Test-Words, Wisconsin Card Sorting Test. Neuropsychological performance was the strongest correlate of domestic violence of all clinical variables measured. However, the inclusion of two other variables, severity of emotional distress and history of head injury, together with the neuropsychological indices provided the strongest correlation with batterers status. Among batterers, neuropsychological performance did not vary as a function of head injury status, indicating that while prior head injury was correlated with batterer status, it was not the sole basis for their impairments. The findings suggest that current cognitive status, prior brain injury, childhood academic problems, as well as psychosocial influences, contribute along with coexisting emotional distress to a propensity for domestic violence.