Methods: The study used data from a non-probability sample of 13 to 19-year-old youths in two Western Pennsylvania long-term residential treatment facilities for juvenile offenders (males = 151, females = 101). Youth were committed to the facilities for various property, person, drug, and status violations. Most youths had years of delinquent behavior and involvement in juvenile court. Violent Behavior (dependent variable, M = 14.04, SD = 12.11) was assessed from a scale asking youth about their rule and law violating behaviors, ranging from a score between 0 and 80. Moral reasoning (independent variable, M = 55.49, SD = 6.74) was measured by a standardized scale including 15 items ranging from 15 to 75. Neighborhood Disorder (M = 15.92, SD = 4.22) was measured by indexing 8 statements relating to the youth’s neighborhood. School Commitment (M = 67.81, SD = 9.91 ) measured how much control the youth perceive they have over their educational outcomes. Scores range from 11 to 77. Emotional Control (M = 12.61, SD = 8.25) was calculated from questions of the Youth Questionnaire Scale. Scores on this scale potentially range from 0 to 30. Multivariate regression was done as a whole and by gender to test whether moral engagement will be significantly associated with lower levels of violent behavior as a whole and among both males and females.
Conclusion: Moral engagement was statistically and negatively associated with violent behavior in both males (b=-.47, p < .01) and females (b= -.43, p < .05). Furthermore, among males, emotional control had a significant positive relationship (b = .412, p = .001), while neighborhood disorder, age, and school commitment were not statistically significant. The R-squared for males was 0.252, indicating 25.2% of the variance in violent behavior among males could be explained by the tested model. Among females, emotional control had a significant positive association (b = .474, p = .001). Whereas like males, age, school commitment, and neighborhood disorder did not significantly predict violent behavior. The R-squared for females was 0.226, indicating the model explains 22.6% of the variance in violent behavior among females. Results were supportive of the hypotheses. The findings suggest that moral engagement and emotional control are crucial in understanding violent behavior.