Can we better align officers' decisions with policy to improve resource use, fairness, and safety?
Focusing on two key decision points—supervision level assignments and technical violation reports—we identify gaps between policy and practice and help design targeted reforms to improve outcomes.
Examining probation officers’ discretion
Every day, probation officers strive to balance two sometimes competing imperatives: keeping communities safe and protecting probationers’ liberty, employment, and family ties. Official risk-need tools and written guidelines are intended to steer these decisions, yet officers retain significant discretionary power. This discretion raises two critical questions. First, do supervision levels and violation responses consistently reflect each person’s likelihood of reoffending? Second, are Black probationers subjected to disproportionately stringent conditions compared to White probationers? A corrections department in a Southern state seeks to understand these decision processes and design reforms that both safeguard communities and promote equitable treatment.
Study objectives
- Measure racial gaps in initial supervision-level assignments and in the filing of major violation reports.
- Assess discretionary overrides, including how often officers override risk-assessment recommendations, and whether overrides disproportionately impact Black probationers.
- Evaluate a policy reform, i.e., the effects of a low-risk, lower-supervision policy on placement decisions, racial fairness, and recidivism rates.
Approach & preliminary findings
We examined state corrections case records (2021-2023). Using multistage statistical models, we adjusted for legally relevant factors such as risk scores and offense types to isolate whether—and to what extent—race influenced an officer’s choice of supervision intensity or decision to file a major violation report. We further traced the frequency and context of instances in which officers chose to override the tool’s recommended supervision level.
Our early analyses revealed that, when we account for risk and offense characteristics, racial disparities in officers’ decisions are surprisingly small. Yet we also identified a clear “decision-support gap”: many individuals assessed as low-risk were nonetheless placed on higher supervision levels than recommended. This mismatch could impose unnecessary constraints on low-risk cases and divert scarce resources from their higher-risk counterparts.
Next steps & potential implications
The state’s corrections agency refined its initial placement guidelines and rolled out targeted training designed to increase officers’ confidence in assigning lower supervision to low-risk cases. We are now evaluating how these reforms affect both placement decisions and recidivism rates. Ultimately, this work aims to demonstrate how carefully tailored policy tweaks and supportive training can help probation systems operate more efficiently, fairly, and safely.
Partners & funding
This evaluation is led by the UC Berkeley Risk Resilience Lab (Jennifer Skeem, Perman Gochyyev, and Sharon Farrell) in close collaboration with the state’s corrections agency. Data sharing and methodological support are provided under a formal agreement with the research division of the state corrections department.