Pretrial Risk Assessments: Garbage in, Garbage out

Jurisdictions Should Avoid Trashing Their Pretrial Release Process with Computer Algorithms

Like a bad bail reform movie sequel, it appears that conversations and proposals involving pretrial risk assessments have once again re-emerged in the bail reform debate.  So far this year we have seen two states, Washington State and South Dakota, submit proposals to incorporate risk assessment solutions into their criminal justice process.  Luckily, South Dakota leadership had the insight and knowledge to stop these faulty pretrial risk assessment efforts in South Dakota in their tracks.  Meanwhile we are waiting to see what happens in Washington State.

While proponents of bail reform will never admit it, there are still many unanswered questions and concerns over the reliability and objectivity of risk assessments to determine pretrial release decisions.  In fact, there are several opinion pieces that have been drafted by leading academic institutions that warn us about the use of these inefficient and discriminatory tools.

An article in the “Knowledge Hub” section of the Roche Surety website, the author not only mentions these academic warnings, but also discusses the discriminatory nature of these flawed instruments.   Leading institutions like Harvard Law, Columbia University, MIT, and Princeton warn us of the critical technical flaws in these risk assessment tools, which fail to accurately distinguish individuals' risk of violence, often producing inaccurate scores due to reliance on flawed, biased data—exemplifying the "garbage in, garbage out" principle. For instance, an analysis of the widely used Public Safety Assessment (PSA) reveals that 92% of those flagged as high-risk for violence did not commit such crimes during pretrial release, leading to excessive false positives that overestimate reoffending and result in higher incarceration rates, undermining the goals of bail reform. Moreover, these risk assessments perpetuate racial disparities in pretrial release by drawing on historically biased arrest and conviction data, where African Americans and Hispanics face disproportionate scrutiny compared to whites for similar conduct, potentially amplifying inequities in the criminal justice system rather than resolving them.  An excerpt from the Roche Surety article is below as well as a link to the full article.

Pretrial Risk Assessments Raise Concerns
by Roche Surety

As more states try to reform the bail system, studies continue to surface highlighting some of the concerns with Pre-Trial risk assessment tools which would determine whether to safely release an individual. Recently, there was a signed statement that was publicly released and endorsed by several prestigious universities, including Harvard Law, Columbia University, MIT, Princeton, among others. The statement noted there are technical flaws that are associated with risk assessment tools that result in the inaccuracy of the scores they generate. READ THE FULL ARTICLE HERE>>>