An Examination of the Effects of Varying Types of Childhood Violence Exposure on the Aggressive Behaviors during Adolescence: Using a Sample of Predominantly Youth of Color from Low-Income Households

Background and Purpose

Childhood violence exposure is known as a risk factor for adolescents’ aggression. However, little is known about if certain types of violence exposure are particularly more influential than others, particularly for marginalized/minoritized groups of youth disproportionately exposed to violence. Based on the assumption of the social-ecological framework (Bronfenbrenner, 1994) that the experiences in proximal systems (e.g., home) are more influential than the experiences in distal systems (e.g., neighborhood) on the individual’s development, this study compared the influence of childhood violence exposure in differing social-ecological systems on adolescent aggression.

Methods

Data were obtained from Wave 5 (i.e., “Year 9”; when the youth was nine years old) and Wave 6 (i.e., “Year 15”; when the youth was 15 years old) from the Future of Families and Child Wellbeing Study. The sample (N = 2,141) was 50.7% boys, predominantly youth of color (46.7% Black, 25.7% Hispanic, 19.9% White), and 55.9% from low-income households. The dependent variable was the youth’s aggressive behaviors at Year 15. The independent variables were caregiver aggression (CA), exposure to intimate partner violence (eIPV), bullying victimization (BV), and exposure to community violence (eCV) at Year 9. OLS multiple regression with hierarchical entry was performed to examine and compare the effects of the violence exposure. Variables were entered in the reverse order of the proximity of the social-environmental system where the violence occurred: eCV (Model 1), BV (Model 2), eIPV (Model 3), and CA (Model 4).

Results

All four models were significant in predicting the youth’s aggressive behaviors at Year 15. eCV remained significant in all models and the second strongest predictor in the final model (β = .065, p < .01). BV also remained significant in models throughout Models 2-4 and was the strongest predictor in the final model (β = .12, p < .01). eIPV was significant when first added to Model 3 but became nonsignificant in the final model. Finally, CA was the third strongest predictor in the final model (β = .064, p < .01).

Conclusions and Implications

Affirming the link between childhood violence exposure and adolescent aggression, this study additionally revealed differing effects of varying types of childhood violence exposure. Violence exposure at a proximal system was not always a stronger predictor than those at more distal systems; BV was the strongest predictor, eCV showed a larger standardized coefficient than CA, and eIPV was not a significant predictor. Future research should explore the factors that create such differential effects across varying types of violence exposure. The findings also warrant a trauma-informed approach to addressing and preventing adolescent aggression and greater attention to developing appropriate intervention strategies for youth exposed to varying types of violence, particularly among marginalized/minoritized groups of youth.