Presenters

Jacob RybakFollow

Academic department under which the project should be listed

Statistics & Analytical Sciences

Faculty Sponsor Name

Susan Mathews Hardy

Project Type

Event

Abstract (300 words maximum)

What leads an offender to go back to prison? This researcher has lived in the Georgia State prison system for 3.5 years. Using personal insights as well as analytics, this researcher analyzes Iowa state’s six-year data set tracking recidivism of released offenders and recommends changes to the prison system to address the analytical findings.

The Iowa recidivism data set includes the following information for all offenders: age group, type of release (parole vs different discharges), release year, original offense, and whether they recidivated. For the recidivating offenders, the data set includes the days to return to prison, the type of recidivism (technicality or new crime), and the offense that caused their return. In the past, the data set has been used with machine learning algorithms to predict whether an offender will recidivate.

The researcher uses logistic regression to determine which variables are best at predicting recidivism. In addition, the interrelationships between variables are investigated with parametric hypothesis tests with post hoc comparisons and graphical displays to convey the findings.

Findings indicate that the original offense predicts the recidivating offense, prompting the question of whether rehabilitation is effective. Another finding indicates that different release types are related to whether an offender recidivates, and the number of days to return to prison, if they do recidivate.

Considering these analyses and others, possible changes are proposed in parole, rehabilitation, and the prison environments. Reforming the prison system is complicated, but if people, especially those who are ex-offenders, join forces to find solutions, more offenders will join this researcher in being rehabilitated and reformed.

Share

COinS
 

Why Does an Ex-Offender Reoffend?

What leads an offender to go back to prison? This researcher has lived in the Georgia State prison system for 3.5 years. Using personal insights as well as analytics, this researcher analyzes Iowa state’s six-year data set tracking recidivism of released offenders and recommends changes to the prison system to address the analytical findings.

The Iowa recidivism data set includes the following information for all offenders: age group, type of release (parole vs different discharges), release year, original offense, and whether they recidivated. For the recidivating offenders, the data set includes the days to return to prison, the type of recidivism (technicality or new crime), and the offense that caused their return. In the past, the data set has been used with machine learning algorithms to predict whether an offender will recidivate.

The researcher uses logistic regression to determine which variables are best at predicting recidivism. In addition, the interrelationships between variables are investigated with parametric hypothesis tests with post hoc comparisons and graphical displays to convey the findings.

Findings indicate that the original offense predicts the recidivating offense, prompting the question of whether rehabilitation is effective. Another finding indicates that different release types are related to whether an offender recidivates, and the number of days to return to prison, if they do recidivate.

Considering these analyses and others, possible changes are proposed in parole, rehabilitation, and the prison environments. Reforming the prison system is complicated, but if people, especially those who are ex-offenders, join forces to find solutions, more offenders will join this researcher in being rehabilitated and reformed.