Abstract
The purpose of this study was to examine the relationship between offender status and various predictor variables with the goal of finding a model of individual characteristics, amenable to change, that could reliably predict violent from non-violent offenders. This study has two main objectives: (a) To determine which personality, behavioral and intellectual variables correlate the strongest with offender status, and (b) to construct a logistic regression model that can help identify which juveniles are more likely to be incarcerated for violent versus non-violent offenses. The participants for this investigation were 95 juvenile offenders referred to the Juvenile Court Assessment Center by the Juvenile Justice Division of the Eleventh Judicial Circuit of Dade County, Florida. This court ordered assessment included the following measures: (a)The Millon Adolescent Clinical Inventory (MALI), (b) the Behavior Assessment System for Children (BASC), (c) the Peabody Picture Vocabulary Test-Third Edition (PPVT-III), (d) the Wide Range Achievement Test Third Edition (WRAT-III), (e) the Kaufman Brief Intelligence Test (K-BIT), and (f) records of school achievement. Out of these measures, the nine variables that had the strongest correlation with offender status were entered into the logistic regression analysis. These nine variables included Reading Percentile, PPVT-III, Inhibited, Eating Dysfunction, Attitude to School, Oppositional, Self Reliance, Peer Insecurity, and Sense of Inadequacy. A logistic regression model comprised of five of these strategically chosen predictor variables (Reading Percentile, PPVT-III, Inhibited, Sense of Inadequacy and Eating Dysfunction) successfully differentiated violent from non-violent juvenile offenders. This model accurately predicted violent from non-violent offenders 86.3% of the time. Specifically, violent offenders were correctly classified 92.2% of the time, while non-violent offenders were correctly classified 74.2% of the time. The variable that was most central in the prediction of offender status was reading percentile. Gender differences were found; however, because of the relatively small sample sizes, the resulting gender specific models should be interpreted with caution.