When battered women with children leave their partners, the professionals who are asked to determine the best interests of the children are faced with a dilemma. They know that the man has a propensity for aggression, which could be directed at the children. Yet, they may also know that the battered woman is also at risk for abusing her children. When making recommendations regarding child custody and visitation, the practitioner must often decide which parent is the least likely to abuse the children and to be the most suited to meet the children's needs.
Social workers may be involved in custody and visitation decisions as child protection workers, family court counselors, mediation counselors, or therapists, as well as expert witnesses (Lyttle-Vieira, 1987). Social workers have the advantage of being trained to conduct person-in-situation assessments and to obtain detailed developmental histories of each parent, assessments that are extremely valuable in custody disputes (Lyttle-Vieria, 1987).
This article reviews the empirical literature on the probability that the man who batters and the battered woman will physically and emotionally abuse the children. Interventions most likely to prevent abuse are presented, and problems with the use of mediation and joint custody are discussed. This article will focus on woman abuse because it is the most common form of spousal violence requiring intervention. Men tend to initiate violence more often and cause more injury (Saunders, 1988; Stets & Straus, 1990). In this article terms referring to married couples also include cohabiting couples, because these couples report rates of violence comparable to those of married couples (Yllo & Straus, 1981). Parent Most at Risk to Abuse
Most studies of domestic violence that also explore child abuse have serious methodological shortcomings. Most use nonrandom samples, do not adequately define abuse, and obtain reports of abuse from only one parent. In domestic violence studies, about half the men who batter are reported to abuse their children (range, 47 percent to 54 percent). Slightly more than a third of the battered women on average reported that they did so (range, 28 percent to 47 percent) (Gayford, 1975; Stark & Flitcraft, 1988; Walker, 1984; Washburn & Frieze, 1980). One study simply stated that battering husbands were usually the abuser of the children, compared with 25 percent of the battered women (Layzer, Goodson, & de Lange, 1986).
Giles-Sims (1985) measured abuse more precisely. Sixty-three percent of the battered women she interviewed said that their partners used violence (actions more severe than a slap, spanking, or shove) against a child in the previous year. Fifty-six percent of the women said that they used the same level of violence against a child. However, the men and women differed greatly in their frequency of violence; the men were reported to have used severe violence an average of 20 times against a child in one year, compared with only four times for the women.
A large-scale study by Walker (1984) provided some evidence that battered women may displace their anger from their partners onto their children. When living with an abusive man, 16 percent of the women said they directed their anger toward their children, and 5 percent expressed this anger with violence. When living with a nonabusive man, only 3 percent of the women said they directed anger toward the children, and less than 1 percent showed their anger with violence. That battered women are not prone to use harsh punishment is supported by a study by Hershorn and Rosenbaum (1985) in which battered women did not differ from two control groups on a standardized measure of punitive child-rearing methods.
The most convincing evidence of differing levels of risk for battered women and their partners comes from the nationally representative survey by Straus (1983). Abuse (as opposed to ordinary punishment) was defined as violence more severe than pushing, grabbing, spanking, slapping, or throwing. …