Co-occurring partner violence and physical child abuse: a test of competing models

In a review of the literature on co-occurring partner violence and physical child abuse, Appel and Holden (1998) found a significantly elevated rate (40%) of child abuse in samples of battered women, compared to the base rate of 6% in representative community samples. Given the considerable overlap between partner violence and child abuse, it is important to develop and test theoretical models that explain the link between these two forms of abuse. The purpose of the present study was to test several possible competing models that vary on identity of the perpetrator and the direction of the violence. In order to do that, the following components within the family were analyzed: direction and intensity of couple violence, externalizing child behavior, maternal parenting stress and antisocial behavior, the co-occurrence of partner violence, child abuse, and the chronological relationships between these components were examined. Participants were 46 victimized women recruited from the community in a medium-sized southwestern police department. Mothers were currently living with their children, and they reported primarily on one of their children aged four to 15 years. Results indicated that target children had been abused by partners in 34.3% of families, and by mothers in 4.3% of families. Retrospective survival analysis revealed that the onset of “more extreme” bi-directional couple violence was significantly related to the onset of partner-to-child abuse. Results indicated that a bi-directional model of co-occurrence most accurately describes violence between family members. The onset of child externalizing behavior was also significantly associated with the onset of partner-to-child abuse; and the onset of maternal parenting stress was significantly associated with the onset of mother-tochild aggression. Results also revealed an approximately simultaneous onset for child externalizing behavior, maternal parenting stress, and child abuse, which supports a coercive family process model. Finally, models of an antisocial behavior trait were not supported. Implications for assessment and understanding the dynamics of family violence are discussed.