Top Common Mistakes or Misunderstandings in Juvenile Risk Assessment.

• Author: Larsen, Rachel

A trend is emerging in some juvenile courtrooms and state legislatures to uncritically accept the latest risk assessment tools and adopt them without a thorough understanding of the tools' strengths and limitations. This trend is so prevalent that there are reports that some judges are making decisions around juvenile dispositions entirely based on a number produced by a juvenile risk assessor without first determining anything about the capacity, quality, or validity of the risk assessment.

With this trend in mind, here is a list of four overlooked issues with risk assessments performed in juvenile court:

1. General risk assessment tools can predict recidivism with moderate accuracy, but they cannot predict whether someone will re-offend with a serious or a minor offense.

   Risk assessment tools are, on average, more accurate than non-tool outcomes based on experience, gut feelings, or even clinical-level psychology skills. That does not mean these tools are omniscient. In fact, they are far from it: a risk assessment tool operating at a 60% accuracy rate is considered acceptable and will not be thrown out. Do you know how well the risk assessment performs in your jurisdiction? If you don't, find the statistics keepers in your jurisdiction and ask for the numbers. After all, if one wants to follow scientifically valid best practices, it is necessary to know how accurate the assessment is in practice.

   The other piece of information we need to know is how likely is it that someone will re-offend, according to the assessment that your jurisdiction uses. For example, what are the chances that someone classified as low risk will re-offend? What are the chances that someone classified as medium risk will re-offend? In one study of the YASI (Youth Assessment and Screening Instrument), the one-year new petition rate based on the pre-screen showed that low-risk offenders reoffended at 9.9%; medium-risk offenders reoffended at 24.9%; and high-risk offenders reoffended at 36.8%. (1) The definition of petition was any new referral to the court with new charges.

   And this is where one runs into a limitation on understanding of how to translate the risk assessment prediction into courtroom practice. One might reasonably say about a juvenile offender in the low-risk category that they do not need to have a courtroom intervention and instead should be diverted. Unnecessary involvement in the criminal justice system is harmful to juveniles, and it should be avoided. (2) That is a well-respected tenant of risk assessment science. But, what if the juvenile in question is charged with murder? Should one look at the low-risk score and release them to the custody of their parents?

   For multiple reasons, prosecutors and judges should not release low-risk juveniles based solely on a risk score if there are indications of danger to the public because of a very serious offense. The first is that low risk may mean that the juvenile has a...