Developing Data-Driven Supervision Protocols for Positive Parole Outcomes

This research will utilize the Georgia Board of Pardons and Paroles' transactional database, which logs every parolee-officer transaction, to identify optimal community supervision strategies. When making supervision decisions, officers rely on the agency's Behavior Response and Adjustment Guide (BRAG), a detailed guide to defining behaviors defined as 'violations or successes' of parole supervision and a range of appropriate officer 'sanctions or responses.' Each violation/success and subsequent response is captured in the automated case management system, which is profiled in a recent NIC technical assistance manual as an example of 'best practices' in parole supervision.

The project will begin with an innovative analysis of the daily parolee-officer transactions of 29,537 parolees to identify temporal and sequential patterns predictive of parole outcomes. The goal is to identify and evaluate case management and supervision strategies that are most effective in achieving successful parole outcomes (defined by multiple measures). The project will culminate in practice protocols that an officer will follow when presented with a parolee violating conditions of supervision or making measurable positive strides toward successful supervision completion. These supervision protocols will be offender specific (based on offender risk), behavior specific (violation/success), and time dependent.

The second phase of the study will evaluate the protocols on a new cohort of parolees as they begin supervision. Parolees will be randomly assigned to experimental and control groups (current practice vs. new protocol), which will dictate the official response of officers to violations or successes. The goal of the evaluation will be to determine if the new protocols for handling parolee violations and successes significantly improve parole outcomes.