5. Mediators of Childhood Exposure to Violence and Intimate Partner Violence
A large problem with the intergenerational violence studies is that too much emphasis is often given to the simple association found, even if weak, and people assume that everyone who had a violent childhood will be violent to their own spouses and children. In fact, researchers have identified both child and spouse abusers who came from nonviolent families and nonviolent individuals who came from violent families. Kaufman and Zigler (1987) reviewed the literature cited to support the intergenerational theory of violence and postulated that the best estimate of the rate of intergenerational transmission was about 30 percent, plus or minus 5 percent. Thus, while approximately one-third of those who have suffered physical or sexual abuse or neglect as children will subject their own children to some form of abuse, two-thirds will not.
Researchers know little about why the majority of abused children do not become violent. Most studies reflect only the linkages between observations of violence and direct experiences with violence during childhood and later behavioral outcomes in adulthood and have not incorporated the intervening variables which ultimately may be responsible for determining whether a person will perform a learned behavior. According to Bandura (1969), exposure to violence does not ensure observational learning. A comprehensive theory of observational learning includes four component processes that influence its nature and degree: attentional processes, retention processes, motor production processes, and incentive and motivational processes. Some people fail to learn the essential features of the model’s behavior, memories may be lost or altered with the passage of time, physical capabilities may restrict performance of a learned observation, and a learned behavior may not be expressed if it holds no functional value for the person or if the behavior is not reinforced. Breakdowns in any of the above processes may result in a failure to translate observational learning to behavior.
Hotaling and Sugarman (1990) updated their earlier review of the literature on intergenerational transmission using multivariate statistics from a national probability sample and could find no link between current marital violence and earlier family-of-origin violence. They now conclude that the relationship that is typically found between current and past childhood violence disappears when other risk factors are controlled, such as socioeconomic status and marital conflict.
There are several potentially confounding social, family, and contextual factors that may be associated with both childhood exposure to violence and increased risks of later adjustment problems or intimate partner violence. Families who experience intimate violence often experience other mental health risks, such as unemployment, drug and alcohol abuse, divorce, incarceration, and other family stressors. Other variables that might mediate the relationship include frequency and duration of exposure, severity of childhood violence, age, gender, perceived legitimacy of violence in family relations, quality of attachment with caregivers, maternal stress, family disadvantage, marital discord, and other stressful life events (Kolbo, Blakely, and Engleman 1996). Other mediators include elevated depression (McCloskey and Lichter 2003) and childhood neglect (Andrews and Brown 1988).
The extent to which confounding factors such as parental criminality, alcoholism, drug use, and adverse life events might explain the relationship between interparental violence in childhood and psychosocial adjustment in young adulthood was examined in a 1977 New Zealand birth cohort of 1,265 children who were followed into adulthood (Fergusson and Horwood 1998). At age eighteen, retrospective reports of interparental violence were obtained. A substantial amount of the association appeared to reflect these social and familial contextual factors. Statistical control of family context was sufficient to explain all or most of several outcomes, including depression, suicide attempts, substance abuse (other than alcohol), nicotine dependence, and violent crime. Associations persisted for anxiety, conduct disorder, alcohol abuse/dependence, and property crime.
It has been suggested that child abuse affects later intimate aggression by enhancing the development of a problem syndrome in adolescence and young adulthood. Using a longitudinal community study, a direct effect between harsh physical punishment in childhood and perpetration of violence against an intimate partner later in life was found. Over half of the effect was indirect through problem behaviors in adolescence and young adulthood (Swinford et al. 2000). In contrast, Mihalic and Elliott (1997) found no direct or indirect effects between self-reported child abuse (‘‘beaten as a child’’) and partner violence later in life.
Path models have demonstrated an indirect path between observing violence as a child and later severe marital violence via sex-role egalitarianism and approval of marital violence, both of which directly influenced the use of severe violence (Stith and Farley 1993). As egalitarianism decreased and approval of marital violence increased, the level of severe violence increased. In this same study, observation of parental violence was also related to decreased self-esteem, which increased the level of alcoholism and marital stress, both of which had an effect on the approval of marital violence. The variables in this study were not measured in temporal sequence; hence no conclusions regarding causality can be made. A national longitudinal study that provided temporal sequencing showed that the path between witnessing parent violence and later partner violence among females was mediated by the development of adolescent delinquency, which resulted in lower marital satisfaction. There was no direct or indirect path between witnessing violence and later partner violence for males (Mihalic and Elliott 1997). Other path models have demonstrated that antisocial personality disorder mediated the effects of abuse/neglect on inter-partner violence for men and women, and hostility and alcohol problems also mediated the effects for abused and neglected women (White and Widom 2003).
A review of the mediating factors that diminish the likelihood of abuse being transmitted across generations suggests that the cycle of violence is less likely to repeat itself if as a child one had the love and support of at least one parent; a loving, supportive relationship as an adult; fewer stressful events in life; and acknowledgment of the childhood abuse and determination not to repeat it (Kaufman and Zigler 1987). Past or current life stresses or supports are influential in determining whether or not the cycle of violence is repeated. Respondents who were not physically abused as children but who abused their own children reported more neglect, more stresses, and less nurturance in the family of origin than those who did not abuse their own children. Abused respondents who did not abuse their children reported fewer stresses in their families of origin than those abused respondents who had abused their own children (Herrenkohl, Herrenkohl, and Toedter 1983).