Chronic Juvenile Offenders: Exploring Risk Factor Models of Recidivism

This study examined the relationship between known risk factors for youthful offenders and rates of recidivism using Poisson regression models. The sample consisted of 564 male and female juvenile offenders referred to the Juvenile Court Assessment Center (JCAC) by the Juvenile Justice Division of the Eleventh Judicial Circuit of Miami-Dade County. First, data from a clinical interview and the administration of the Wide Range Achievement Test were factor analyzed. Six factors were found to be statistically significant based on a parallel analyses. Neighborhood factors explained the largest amount of variance followed by peer influence, family functioning, gang involvement, substance use, and academic achievement. These six domains were analyzed in separate Poisson regression models. Family-wise error rate was controlled with Bonferroni adjustments. Each model predicting number of arrests from academic performance, substance use, peer influence, gang involvement, and neighborhood factors were statistically significant. The final model including all variables across the six domains indicated good fit, (2)(14) = 201.260, p < .001. Implications stemming from these findings are discussed.