Recidivism, the tendency to relapse to old ingrained patterns of behavior, is a well-documented problem among persons in intimate partner violence treatment programs. In "Pattern of Reassault in Batterer Programs" (Violence and Victims, vol. 12, no. 4, 1997), Edward Gondolf reported his evaluation of four well-established batterer programs to assess the pattern of reassault or a return to battering. All the research sites had operated for five years or more and received at least forty to fifty referrals per month. Located in Pittsburgh, Pennsylvania; Denver, Colorado; and Dallas and Houston, Texas, the programs ranged from three to nine months in duration.
Of the 840 batterers recruited, 210 at each site, 82% were referred to the program by court order, and 18% entered the program voluntarily. Both batterers and their partners were interviewed by phone every three months for fifteen months after intake. The female partners of 79% of the batterers were interviewed at least once during the fifteen-month follow-up.
In follow-up reports on 662 batterers, 32% of the female partners reported at least one reassault during the fifteen months after treatment. Of the 210 reassault cases, 61% resulted in bruises or injuries, and 12% of victims required medical attention. The reassault rate was significantly higher for program dropouts than for participants who completed the program. Voluntary participants were also more likely to reassault their partners than court-ordered participants.
While the proportion of women who were reassaulted was relatively low, 70% of the women were subjected to verbal abuse, 45% were subjected to controlling behaviors, and 43% experienced threats. Nonetheless, 66% of the women said their "quality of life" had improved, and 73% reported feeling "very safe" during the follow-up periods.
Fourteen percent of first-time reassaults occurred in the first three months of the program, and 8% occurred within four to six months. Early reassault appeared to be a high-risk marker for continued abuse. Men who reassaulted their partners within the first three months were much more likely to repeat their attacks than were men who reassaulted for the first time after the first three months. The repeat offenders were also highly likely to use severe tactics and inflict injuries. Gondolf speculated that intervention may have been less effective for this group of men because of previous contact with the criminal justice system and/or severe psychological disorders.
Gondolf concluded that well-established programs seem to contribute to the cessation of assault, at least in the short term. For "resistant batterers," he recommended more extensive monitoring and intervention.
Saunders, in "Husbands Who Assault: Multiple Profiles Requiring Multiple Responses," reviewed the available information on male batterers and found that the recurrence of violence six months or more after treatment averages 35% across a number of studies (Legal Response to Wife Assault, [Newbury Park, CA: Sage, 1993]). For men who do not complete treatment, the average reassault rate is 52%. The men most likely to return to violence are on average younger, report alcohol problems, score higher for narcissism (excessive self-involvement) on psychological tests, and have longer histories of pretreatment violence.
Julia C. Babcock and Ramalina Steiner reported some cautiously optimistic findings in "The Relationship between Treatment, Incarceration, and Recidivism of Battering: A Program Evaluation of Seattle's Coordinated Community Response to Domestic Violence" (Journal of Family Psychology, vol. 13, no. 1, March 1999). Their research measured recidivism of domestic violence after arrest and completion or noncompletion of a mandatory, coordinated program of treatment involving the courts, probation officers, and treatment providers.
Babcock and Steiner followed 387 people arrested for misdemeanor domestic violence offences, thirty-one of whom were women. More than three-quarters of participants had no prior domestic violence convictions and 69% had no prior criminal history. The average age of participants was 32.7 years, 45% had graduated from high school, and 36% had attended college or were college graduates. About 41% of participants were white, 36% were African American, 6.6% were Hispanic, 8.6% were Asian American, and 7.8% identified themselves as "other." Half of the participants were employed and 31% were married.
Participants were referred to one of eleven certified domestic violence treatment programs. The majority attended programs that use the Duluth model, while the remainder participated in feminist, psychoeducational, and cognitive-behavioral men's groups. About 31% completed at least twenty-four sessions of treatment, and those batterers considered to have completed treatment attended an average of thirty-two sessions. In contrast, batterers who did not complete treatment attended an average of just 5.8 sessions. Treatment completers were generally first-time offenders, better educated, employed, and had less prior criminal involvement. Of the noncompleters, 58% did not attend any sessions, but the majority were not legally punished, despite their failure to attend court-ordered treatment.
Program completion was related to lower rates of recidivism—treatment completers had significantly fewer domestic violence arrests at follow-up than noncompleters, and this difference remained even when the researchers controlled for differences in prior criminal record and history. Batterers who had been court ordered to attend treatment and failed to complete it were more likely to commit further offenses than treatment completers. Babcock and Steiner concluded that their findings support the premise that completing treatment is directly related to reduced rates of domestic violence. They cautioned, however, that participants who completed treatment were probably not representative of the entire population of batterers—they likely had more to lose as a result of failure to complete treatment than the treatment dropouts.
Jill A. Gordon and Laura J. Moriarty in their study of the effect of batterer treatment on recidivism reported more pessimistic results ("The Effects of Domestic Violence Batterer Treatment on Domestic Violence Recidivism," Criminal Justice and Behavior, vol. 30, February 2003). They found that attending treatment had no impact on recidivism when comparing the treatment group as a whole with the experimental group. However, they also found that among the treatment group, the more sessions a batterer completed, the less likely he was to batter again. Batterers who completed all sessions were less likely to be rearrested for domestic violence than were batterers who had not completed all sessions.