Once someone has been convicted and sentenced for a crime, corrections agencies use risk and need assessment (RNA) tools to identify how likely that person is to commit another crime or violate the rules of prison, jail, or community supervision. Correctional authorities use RNA instruments to guide decisions about programming, support, and restrictions that are intended to enhance public safety and make better use of scarce resources. Over the past several decades, the use of RNA in correctional systems has proliferated. Indeed, the vast majority of local, state, and federal correctional systems in the United States now use some type of RNA. Despite the numerous ways in which RNA instruments can improve correctional policy and practice, the kind of RNA currently used across much of the country has yet to live up to this promise because it is outdated, inefficient, and less effective than it should be.
In an effort to help the corrections field realize the full potential of RNA instruments, NIJ recently released Guidelines for Post-Sentencing Risk Assessment.[1] These guidelines, assembled by a trio of corrections researchers and practitioners, are built around four fundamental principles for the responsible and ethical use of RNAs: fairness, efficiency, effectiveness, and communication. Each of these principles contributes to an innovative, practical checklist of steps practitioners can use to maximize the reliability and validity of RNA instruments.
The first principle, fairness, holds that RNA tools should be used to yield more equitable outcomes. When designing risk assessments, we need to address and overcome potential sources of bias. Accounting for the possibility of bias from the beginning can help shape RNAs to mitigate racial and ethnic disparities in the way people are assessed, rather than perpetuating those disparities. Disparities can also be reduced by shifting the way RNAs are used, such as delivering more programming resources to those who need them the most. Designing for fairness provides correctional agencies with a strategy for achieving better, more just outcomes.
The second principle, efficiency, indicates that RNA instruments are more reliable when they are more automated. Right now, most RNAs involve scoring each person manually on items like criminal history, demographic characteristics, dynamic risk factors, and program participation. Running an RNA on manual inputs takes a great deal of time, and it also opens the door to subjective and biased scoring. Computer-assisted scoring, in contrast, ensures that an RNA’s inputs are more consistent and reliable, which is more fair and also takes up less time and fewer resources.
RNA instruments should not only be fair and efficient; they also need to be effective. Significant advances in statistics, data science, and predictive analytics have introduced new options for RNA tools that make better predictions. For example, machine learning algorithms offer increased predictive accuracy and overcome some of the limitations of older statistical techniques. It is also important to customize RNA tools to the population on which they will be used. Instruments built and tested with local data will perform better than off-the-shelf instruments designed for incarcerated persons from another state or jurisdiction.
Finally, beyond making technical improvements in the way RNAs are designed and applied, it is just as important to focus on their implementation. Assessing individuals’ risk must go hand in hand with increasing their awareness of the risk factors that apply to them. To this end, improved communication is the fourth key principle. We need to train correctional staff so they can explain risks and needs and translate them into a case plan with clear sanctions and incentives. Effective communication helps individuals understand how their risks and needs can be addressed with programming, empowering them to work on the factors that affect their involvement in crime. Sharing assessment information as a matter of agency policy also promotes fairness and transparency.
Reliance on these principles will produce RNA tools that reduce disparities and achieve better recidivism outcomes. Taking advantage of improved methods for designing these tools can help corrections practitioners around the country build RNA systems that are both more efficient to operate and fairer to those they assess. Still, these approaches are new, innovative, and not yet widespread, opening the way for significant transformation of the corrections field in the years to come. In the words of the guidelines’ authors, “We believe that full implementation would lead not only to more responsible and ethical use of RNA tools but also to better, more equitable outcomes for correctional populations and systems.”
About This Article
The research described in this article was funded by NIJ contact 2010F_10097. This article is based on the white paper, Guidelines for Post-Sentencing Risk Assessment (July 2021) by Kristofer Bret Bucklen, Grant Duwe, and Faye S. Taxman.