|dc.description.abstract||This dissertation investigated the stability of race used as a predictor of recidivism on African American prisoners in South Carolina who received career and technical education while incarcerated. The purpose was to determine whether variables of security class, number of disciplinary reports, age at release, education level at intake, crime type, number of prior incarcerations, race, sex, and sentence length predicted recidivism for adults incarcerated in South Carolina who completed a career and technical education certificate program while in prison and those who did not completed a correctional education program.
Data from all prisoners released between January 1, 2004 and December 31, 2005 were included. The selected variables were used to create a model to predict recidivism in two samples: prisoners who had completed a career and technical education certificate while incarcerated and prisoners who had not completed an education certificate while incarcerated. This was done to both determine if these variables were statistically significant in predicting recidivism, as well as to determine if the same variables were significant in both populations.
Stepwise logistic regression was conducted on both samples to determine the best model, age at release, number of prior incarcerations, education level at intake, security class and crime type were found to be significant predictors of recidivism for prisoners who completed a career and technical education certificate. Age at release, number of prior convictions, education level at intake, security class, crime type, race, sex and number of disciplinary reports were significant in the non-completer sample. Both models predicted recidivism more effectively than chance.||