Technical Flaws of Pretrial Risk Assessments; 27 Leading Researchers Raise Grave Concerns

In July of 2019, twenty-seven of the nation’s leading researchers in the area of predictive algorithms sent letters of concern to several political leaders in California and Missouri.  These researchers represented a broad spectrum of academic institutions including Harvard, MIT, Princeton, NYU, Columbia and UC Berkeley. These letters outlined the concern that these researchers had about the growing support for the use of algorithms in the criminal justice system, specifically in California and Missouri.  Both states have recently planned to make wholesale changes to their criminal justice systems by incorporating the use of risk assessments into their pretrial processes.  The correspondence from these 27 researchers was intended as a warning to these jurisdictions to beware of these algorithms.

According to the research group, these algorithms have fundamental flaws that will lead to negative consequences.  They stated, “Although pretrial risk assessment tools are often promoted as an essential part of bail reform that can help judges make more informed and objective pretrial decisions, these tools suffer from serious methodological flaws that undermine their accuracy and effectiveness. As a result, pretrial risk assessments do not increase the likelihood of better pretrial outcomes, much less guarantee them.”

Accompanying the letter was a research report entitled, “Technical Flaws of Pretrial Risk Assessments Raise Grave Concerns.”  The report concluded that “Pretrial risk assessments do not guarantee or even increase the likelihood of better pretrial outcomes. Risk assessment tools can simply shift or obscure problems with current pretrial practices.”

The 27 research scientists conclude, “We strongly recommend turning to other reforms.”

You can read the press release by clicking HERE.

You can read the letter of concern in it’s entirety by clicking HERE.