Beliefs as Mediators of Relations Between Exposure to Violence and Physical Aggression During Early Adolescence

Abstract

This study examined beliefs about aggression and self-efficacy for nonviolent responses as mediators of longitudinal relations between exposure to violence and physical aggression. Participants were a predominantly African American (79%) sample of 2,705 early adolescents from three middle schools within urban neighborhoods with high rates of violence. Participants completed measures across four waves (fall, winter, spring, and summer) within a school year. Beliefs supporting proactive aggression, beliefs against fighting, and self-efficacy for nonviolence partially mediated relations between witnessing violence and physical aggression. Indirect effects for beliefs supporting proactive aggression and self-efficacy were maintained after controlling for victimization and negative life events. Beliefs supporting proactive aggression mediated the effects of violent victimization on physical aggression, but these effects were not significant after controlling for witnessing violence and negative life events. The findings underscore the importance of examining the unique pathways from witnessing community violence versus violent victimization to physical aggression.

Keywords: African American adolescents; Aggression; Exposure to violence; beliefs about aggression; self-efficacy.

Figure 1:

Standardized path coefficients for one-sided model representing self-efficacy for nonviolence as a mediator of the relation between witnessing violence and changes in physical aggression. Paths represented by dashed lines were included in the model, but were not significant at p < .05. Effects of covariates (sex, grade, and intervention status) on each variable, and correlations among residuals for variables within the same wave were included in the model, but are not shown. Witnessing violence was treated as an exogeneous variable and was allowed to correlate with itself across waves, and with self-efficacy for nonviolence and physical aggression at the same wave and at all prior waves.

Figure 2:

Standardized path coefficients for one-sided model representing belief variables as mediators of the relations between negative life events, victimization, and witnessing violence on changes in physical aggression. Paths represented by dashed lines were included in the model, but were not significant at p < .05. Effects of covariates (sex, grade, and intervention status) on each variable, and correlations among residuals for variables within the same wave were included in the model, but are not shown. The three exposure variables were treated as exogenous variables and were allowed to correlate with each other across all waves and with belief variables and physical aggression at the same wave and all prior waves.