Bridging Disciplinary Boundaries (January 11 - 14, 2007)

PURPOSE OF THE STUDY. The police have actively intervened in domestic violence through pro-arrest policies since the 1980s, but research results of the effectiveness of arrest are inconclusive. Some insisted that arrest has effectively reduced the domestic violence incidents as shown by the reduced recidivism rates. Others contended that arrest has rather increased victims' risk of revictimization. However, the majority of previous studies failed to yield valid and generalizable results due to their methodological shortcomings, such as non-random small convenient samples and short follow-up periods. This study attempted to overcome such shortcomings by using the National Crime Victimization Survey (NCVS) from 1987 to 2003, to examine the effectiveness of arrest of batterers in reducing victims' revictimization.

METHOD. The NCVS annually gathers detailed crime data from a nationally representative sample of 50,000 households with 100,000 individuals. The target population of the study is the domestic violence victim age 18 and over. A total of 2,462 women were identified as victims and included in the analysis. Logistic regression analysis was conducted, including one dichotomous dependent variable (revictimization), one dichotomous independent variable (arrest), and eight control variables (age, race, income, marital status, number of children, nature of violence, use of a weapon, and injury).

RESULTS. Sample characteristics showed that almost one quarter of batterers (24.2%) in domestic violence cases were arrested, while less than one fifth of victims (16.4%) were revictimized, mostly (92.3%) within 6 months after the previous victimization. Arrest, age, marital status, nature of violence, and injury were shown to have significantly affected revictimization. First of all, arrest of batterers had an effect on preventing revictimization. Victims whose partners had been arrested were less revictimized than those whose partners had not been arrested (exp(β) = .61). Aging had a very small preventive effect on revictimization, but too trivial to be accepted as evidence of effect (exp(β) = .99). Also, separated women were more revictimized than those who have never been married (exp(β) = 1.52). Victims of rape or sexual assault showed the highest risk of revictimization (exp(β) = 3.11), compared with those of simple assault. Finally, victims who had suffered severe injury from the previous offense were less revictimized than those without any injury (exp(β) = 0.64).

IMPLICATIONS. The study results showed that victims whose batterers had been arrested were less revictimized than those whose batterers were not arrested, and the victim's chance of revictimization varied, depending upon sociodemographic characteristics of the victim and the nature of the domestic violence incident. It is necessary for future research to further investigate why and how individual and circumstantial factors affect the chance of revictimization, and how these factors can be integrated into comprehensive understanding of dynamics between police involvement and victims' future safety. Also, it is necessary to develop different police intervention strategies for different victims. Finally, social work practice needs to tailor its services to individual victims' unique situation and needs (e.g., victims separated, sexual assault victims, and victims with severe injury) rather than relying on any one-size-fits-all solution.