Could an Algorithm Help Solve Prison Overcrowding?

Iowa’s prisons are 22% over capacity and have been operating beyond their limits since 1989. The University of Nebraska at Omaha (UNO) and The University of Colorado Boulder (CU Boulder) researchers think an algorithm might be part of the solution.

➡️ What’s new:

• A UNO and CU Boulder research team has received a Arnold Ventures (AV) grant to develop and test a new parole decision-making tool that could help address decades-long prison overcrowding.
• Iowa’s prison system has operated above capacity since 1989 and is currently 22% over capacity. State officials set a goal to reduce the prison population to 100% of capacity by December 2027, and the tool will assist by identifying low-risk individuals for earlier release.
• The tool’s impact will be measured in a year-long field trial: over the course of 12 months, the Iowa Board of Parole will randomly receive the tool’s “yes”/“no” release recommendations for only a subset of eligible cases, allowing researchers to observe how much the recommendations change parole decisions and subsequent recidivism.

💡 Why it matters:

• U.S. prison populations have increased by 700% since the 1970s, straining state budgets and communities. Iowa’s overcrowding is a microcosm of a nationwide challenge. The United States imprisons a larger share of its population than any other independent democracy.
• Releasing too few prisoners worsens overcrowding, but releasing too many or the wrong individuals can threaten public safety. Parole boards face intense pressure to balance these concerns.
• If a risk assessment tool can safely pinpoint who is ready for release, states could return low-risk individuals to society, reduce correctional costs, decrease violence and stress inside prisons, and create a model to ease crowding elsewhere.

🔎 Zoom in:

• The research team is building the tool using years of administrative data from the Iowa Department of Corrections and Judicial Branch, employing a rigorous risk-needs assessment (RNA) algorithm.
• The model sets two risk thresholds – one for any re-offense and one for violent re-offense – and only inmates rated low risk on both receive a “yes” recommendation for early release.
• Each eligible parole case is distilled to a simple “yes” or “no” recommendation. Every month, the lowest risk of reoffenders are identified for potential release, in numbers calibrated to gradually bring Iowa’s prison population down to 100% capacity by the end of 2027.
• Nebraska’s Board of Parole is also observing the project as a potential model. UNO researchers are working with Nebraska officials on a pilot to adapt the tool for Nebraska’s needs, highlighting its broader promise beyond Iowa.

🎤 What they’re saying:

• Jennifer Tostlebe, Ph.D., Assistant Professor of Criminology and Criminal Justice at UNO and project co-investigator: “This collaboration between universities and parole authorities allows us to tailor the tool to real-world needs. By working together with officials in Iowa, we can ensure the science translates into actual safety and policy improvements.”
• Alex Kigerl, Ph.D., Research Associate at UNO’s Nebraska Center for Justice Research and project researcher: “If this works in Iowa, it could be a blueprint for other states. Our goal is a reusable model – a parole decision tool that can help any system facing overcrowding to safely reduce their prison population.”
• David Pyrooz, Ph.D., Associate Professor of Sociology at CU Boulder and co-principal investigator: “We’re conducting a rigorous test of the tool’s impact. By randomly providing recommendations in some cases and not others, we’ll get clear evidence on how much these data-driven insights influence parole decisions and outcomes.”
• Stephen Billings, Ph.D., Associate Professor at CU Boulder and project researcher: “Prison overcrowding is a complex public policy challenge. This project shows how innovation and evidence can inform that challenge – using analytics to enhance decision-making in a way that benefits public safety, individuals, and taxpayers alike.”
• Zachary Hamilton, Ph.D., Professor of Criminology and Criminal Justice at UNO and the study’s principal investigator: “We’re leveraging data in a new way to help parole boards make more informed decisions. It’s about maintaining public safety while finally tackling a prison overcrowding problem that has persisted for decades.”

🤝 Grant information:

• Funding source: Arnold Ventures
• Award number: NuRamp number NRF167191
• Title: Development and Evaluation of a Parole Decision Making Tool
• Note: This material is based upon work supported by Arnold Ventures under Award No. NRF167191. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of Arnold Ventures.