Risk Assessment in the News

October 25, 2013 | Erin Manske, Researcher, NCCD

manskepic

A recent Wall Street Journal article discusses the growing acceptance of the use of risk assessment in parole decisions. However, risk assessments are not new to the criminal justice world. NCCD has been developing actuarial risk assessments for use in juvenile and adult justice settings since the early 1980s. These risk assessments are used by parole and probation departments to guide decisions related to supervisory strategies.

A recent Wall Street Journal article discusses the growing acceptance of the use of risk assessment in parole decisions. However, risk assessments are not new to the criminal justice world. NCCD has been developing actuarial risk assessments for use in juvenile and adult justice settings since the early 1980s. These risk assessments are used by parole and probation departments to guide decisions related to supervisory strategies.

Typically, risk assessments are used to classify people into groups (e.g., low, medium, or high risk) based on a set of characteristics, or “risk factors.” The goal of a risk assessment in a justice-system setting is to identify cases most likely to be involved in future offending. Agencies can use risk classification to target their resources toward cases that are more likely to reoffend. This approach allows agencies to maximize limited resources, such as available staff time to supervise probationers.

Actuarial risk assessment is by no means a new practice, either; this type of risk assessment is based on statistical relationships between individual factors or characteristics and a set of outcomes (e.g., subsequent offenses). To develop an actuarial assessment, large datasets are collected and analyzed. A combination of analyses reveals which factors have the strongest statistical relationships with the recurrence outcomes (e.g., new offense, new adjudication). From here, different combinations of these factors are identified and examined to create a risk assessment that most effectively separates a given population into high-, medium-, and low-risk groups. For example, cases classified as low risk should have lower recurrence rates than cases classified as medium risk, and cases classified as medium risk should have lower recurrence rates than cases classified as high risk. Not only should outcomes follow this “stepwise” pattern in outcomes as risk level increases, but the assessment should also classify a reasonable number of cases into each category—not everyone should classify as low  or high risk.  

NCCD uses a comprehensive, evidence-based, actuarial approach when developing risk assessments. We emphasize that actuarial risk assessments are meant to guide, not replace, the professional judgment of corrections officials. In fact, many risk assessments have an override option allowing corrections officials to have final authority over the risk decision. We also stress that risk assessment is not the same as prediction. For instance, if the recurrence rate for the high-risk group is 60%, we know that 40% of the cases classified as high-risk will not reoffend, but we can’t tell which cases those will be. To ensure we are getting the best results possible, NCCD risk assessments are tested to ensure racial and gender equity, where data are available.

For more information on NCCD’s work in risk assessment, contact Dr. Jesse Russell, Director of Research–Midwest, at (800) 306-6223.

Erin Manske is a Researcher at NCCD.