Back-End Sentencing: An Exploratory Study of Predictors of Parole Revocation

• In Event: Parole and Community Supervision

Abstract

Empirical explorations about the rise of mass incarceration have, with few exceptions, focused on “front-end” policies and practices such as the rise of harsher criminal penalties, the expanded use of mandatory minimum sentences, and truth-in and determinate sentencing policies. However, the process of returning parolees back to prison through supervision revocations, a process commonly known as “back-end” sentencing, has become increasingly recognized as an important contributor to mass incarceration in the United States. The current study adds to the growing knowledgebase about the process of back-end sentencing through the analysis of data highlighting all parole revocation decisions that occurred in New Jersey from 2005 to 2011 (n=13,121). In addition to demographic and case controls, we include key organizational variables to construct nested logistic regression models in an attempt to isolate the revocation decision-making practices of state parole board members. Results show that the “usual suspects” such as age, gender, race, and criminal history variables demonstrate little explanatory power, while actuarial risk levels, programmatic resources geared towards parolees during the course of supervision, and the timing of revocation decisions in relationship to the release and maximum sentence expiration dates provide the greatest predictive strength.

Authors

• Michael Ostermann, Rutgers University

• Jordan Hyatt, Drexel University