Judges can consider predictive data on risk of reoffending in limited circumstances, court rules
Judges can’t use predictive data on an offender’s risk of committing more crimes in determining whether he or she gets prison time or a lengthier sentence, the Wisconsin Supreme Court has ruled.
But predictive data can be used by judges, the court said, as one factor: in identifying prison-bound offenders who could be diverted to alternatives outside of prison, in assessing the safety of supervising an offender in the community, and in imposing terms of probation and supervision. The Wall Street Journal Law Blog has a story on the July 13 opinion.
The court ruled in the case of Eric Loomis, who pleaded guilty to operating a vehicle without the owner’s consent and attempting to flee an officer. The state had maintained Loomis had driven the car in a drive-by shooting; Loomis denied involvement and maintained he drove the car after the shooting occurred.
A judge who sentenced Loomis to six years in prison cited a so-called Compas report that found him to be a high risk to the community. The test makes a prediction based on a comparison of the individual’s information to that of individuals in a similar group.
Loomis asserted that the use of the Compas risk assessment violated his right to due process, either because its proprietary nature prevents defendants from challenging its validity, or because the test takes gender into account.
The Wisconsin Supreme Court found no due process violation because the sentencing judge relied on other, independent factors and said the sentence would have been the same without the Compas report.
The state supreme court also said the test’s use of gender promotes accuracy because men, on average, have higher rates of reoffending and violent crime than women.
Though Loomis couldn’treview how the Compas algorithm calculates risk, he did know that the scores are based largely on criminal history, along with variables such as criminal associates and substance abuse, the state supreme court said. His Compas report lists his answers to 21 questions on criminal history and he had the opportunity to verify the accuracy of the answers.
The supreme court listed information that must be provided to courts considering Compas risk assessments. They should be informed about the proprietary nature of the data, that it compares defendants to a national sample but it has not been tested against a sample of the Wisconsin populations, that some studies have questioned whether minority offenders are disproportionately classified as having a higher risk, and risk assessment tools must be constantly re-evaluated due to changing populations.
Related article:
ABAJournal.com: “Legality of using predictive data to determine sentences challenged in Wisconsin Supreme Court case”
