How Does Child Maltreatment Beget Youth Violence? An Examination of Mixed-Gender and Gender-Specific Mediator Models

Investigators concerned about the cycle of violence have generated substantial evidence linking child physical abuse to adolescent or young adult violence. Child sexual abuse and neglect, however, have also been linked to later violence. Moreover, the association between child abuse and neglect (CAN) and youth violence appears to persist across genders. Given the increase in violent delinquency among females, the level of male youth violence, and the attendant costs of both CAN and violent offending, the longitudinal relation between these two phenomena warrants ongoing attention.

While relatively confident in the main-effect association between childhood victimization and later violence perpetration, we know little about the mediators of this connection. Criminological and developmental theories identify potential mediating pathways, yet these processes have not been well-tested. Therefore, we analyzed hypothesized mediators of the CAN-violence nexus, addressing the following research questions. What processes mediate the association between CAN and adolescent or young adult violent offending? How do these mediating mechanisms differ by gender?

Data derived from the Chicago Longitudinal Study (CLS), a panel investigation of 1,539 low-income, minority participants who graduated from kindergarten in 1986. Prospectively administered assessments along with ongoing administrative record searches yielded measures spanning childhood, adolescence, and early adulthood (ages 18-24). Files from the Illinois Department of Child and Family Services and the Cook County Juvenile Court informed official indicators of CAN and violent juvenile delinquency. County, state, and federal records indicated adult violent crime. Late childhood and early adolescent mediators originated from self-, teacher-and school-reports.

Because prior CLS analyses uncovered a significant main effect relation between CAN and youth violence (Mersky & Reynolds, 2007), we focused solely on tests of mediation. Employing multivariate hierarchical probit regression, we found that the following processes partially explained the association between CAN and any juvenile or young adult violent offending: trouble-making behavior, parental expectation of school success, mobility, and magnet school attendance. Segregating the sample by gender revealed that these same four mediators, along with social-emotional adjustment, fully explained the main-effect relation in question for males. For females, only acting out behavior in school helped to actualize the CAN-violence connection.

Structural equation modeling in LISREL confirmed all mediator models. For each sample specification, the data fit the model well. Although the results for the full and female samples reflected partial mediation, the male mediation model reduced the original main effect relation to statistical non-significance and by a magnitude of 72.6%. Two-group SEM analysis revealed that the male and female mediation models differed significantly.

Results articulate processes that potentiate or ameliorate maltreatment's violence-provoking tendencies, thereby informing interventions. Practitioners working with boys who have been maltreated might enhance the child's social-emotional skills, target caregiver expectation of child's school performance, and collaborate with school personnel. Reducing mobility and trouble-making behavior is also indicated. For girls who have been maltreated, study findings recommend resolution of school-based acting out behavior before it translates into violence. Additionally, our results strengthen the case for integrating child welfare and juvenile justice services.