Participants were 184 children, ages 7 to 13, and their mothers, who were residing at battered women shelters. The Physical Aggression subscale of the Conflict Tactic Scale measured child abuse, and children's behavioral problems were assessed with the CBCL (Achenbach and Edelbrock, 1986). The Physical Aggression subscale of the Conflict Tactics Scale measured the amount of spousal violence; and a single-item measure, correlating at a level of .84 with the 25-item Hudson scale (1982), assessed the quality of parent-child relationships. The short Marital Adjustment Test (Locke and Wallace, 1959) measured marital satisfaction, and alcohol/drug abuse of the parents was determined by asking mothers to report the extent to which alcohol/drug abuse was a problem in their relationship with their spouse/partner. Other variables measured were family size, family socioeconomic status, child temperament, child aggression, and formal and informal social support systems. Findings show that compared to children who witnessed interparental violence, children who were both witnesses and victims of physical abuse scored significantly higher on externalizing behavioral problems. A stepwise logistical regression analysis identified numerous predictors of child abuse among maritally violent families. These included a greater frequency and severity of marital violence, less marital satisfaction, and a poorer quality of father-child relationship as perceived by the children. In addition, children from families where child abuse occurred were more likely to exhibit aggressive behavioral problems than children who only witnessed marital violence. 1 figure, 4 tables, and 72 references