Can structured tools improve the consistency and fairness of parole decisions?
We are analyzing parole processes and outcomes in California to identify any disparities and explore whether data-driven frameworks could deliver fairer, more consistent releases.
Why parole decision-making needs a closer look
Parole boards wield broad authority over how long people serve in prison—yet the processes guiding their choices are largely opaque. In California, more than one-third of the prison population faces discretionary parole review, but we know little about how race or socioeconomic status affects who gets a hearing, who proceeds, or who is ultimately released. At the same time, California has adopted standardized risk assessments and structured decision-making frameworks for parole. This project leverages a unique partnership with the Board of Parole Hearings (BPH) to shine a light on parole practices and explore whether these decision-support tools could reduce any racial or socioeconomic disparities without compromising public safety.
Study objectives
- Map racial and socioeconomic differences at each stage of the parole process—who is scheduled for a hearing, who proceeds to a hearing, and who is granted release.
- Evaluate current use of the BPH’s risk assessment tool and structured decision-making framework, and explore whether stronger adherence could narrow disparities.
- Simulate a “low-risk” release policy—anchored in risk scores—and compare overall release rates and racial equity outcomes to actual practice.
Study approach
Quantitative Mapping of Any Disparities: We will examine administrative records for all 21,000+ parole hearings in California (2021–2023). Multistage statistical models will track selection into hearings—accounting for inmate waivers, postponements, etc.—and then model release outcomes. At each decision point, we will quantify the role of race or socioeconomic status after adjusting for legally relevant factors like institutional behavior.
Qualitative Examination of Decision-Support Tools: In 2019, BPH adopted a structured decision-making framework anchored by a risk assessment. Because framework variables are not yet captured digitally, we will observe parole hearings and qualitatively code a representative, race- or SES-stratified samples of transcripts. We will identify how and when commissioners reference the framework, how it shapes deliberations, and how departures from risk recommendations are justified, paying special attention to any race- or SES-linked patterns in overrides.
Policy Simulation: We will quantitatively examine how closely actual release decisions align with risk-assessment categories (low, moderate, high) across racial or SES groups. We will then compare actual release rates to a modeled presumptive low-risk policy—where low-risk individuals are released by default—to estimate potential gains in fairness and consistency. This will illuminate the potential for structured, score-based policies to increase releases of low-risk individuals and narrow racial gaps in parole decisions.
Next steps & potential implications
When analyses are complete, we will share the results through presentations, policy briefs, and articles. The results of this study will help BPH understand the fairness of current decision-making and the extent to which greater adherence to decision aids could affect the consistency and equity of decisions.
Partners & funding
This independent study is co-led by Jennifer Skeem and Steven Raphael, in collaboration with California’s Board of Parole Hearings. Sharon Farrell leads the qualitative component. UC Berkeley’s effort is supported by a grant from the Crankstart Foundation.
