Courts move forward on pretrial assessment tools after voters OK bail

After voters rejected a measure that sought to replace the state’s cash bail system with a computer generated risk assessment system, the California Judicial Council will hear recommendations this week on how it can improve the algorithmic tools judges in 49 counties already use alongside bail.

The recommendations from a work group formed by the chief justice include: eliminating a defendant’s mental health status from being included as a risk factor, increased training and education for judges to better understand how the tools generate results and continuing research to learn how the tools can be used to identify and mitigate any racial bias in pretrial decision making.

Defense attorneys who opposed Proposition 25 on the Nov. 3 ballot, which would have eliminated cash bail and required courts to uniformly rely on risk assessment tools, have argued that the tools often produce flawed results that keep more people behind bars than cash bail does.

“These are dangerous, inaccurate and racially biased ZIP code driven guides that the courts must discard if real progress is to be made,” said Eric Schweitzer, president of California Attorneys for Criminal Justice, a statewide group of private and public defense attorneys.

Chief Justice Tani G. Cantil-Sakauye appointed the Pretrial Reform and Operations Workgroup in January 2019 to analyze the efficacy of California’s pretrial system and provide recommendations for change. The group is led by 4th District Court of Appeal Justice Marsha G. Slough.

The decision to form the group came after Gov. Gavin Newsom allocated $75 million in his 2019-2020 budget to fund a proposed two-year court pretrial program. The Judicial Council pretrial work group selected 16 courts to pilot the program, which the group analyzed over the last year.

In a near 40-page report to be presented during the Judicial Council’s Nov. 13 meeting, the 12-member group comprised of trial court judges, appellate justices and court executive officers, found that risk assessment tools can be “a valuable addition to a judge’s pretrial release toolbox,” if some changes are made to how they function.

Risk assessment tools currently rely on a defendant’s criminal history, demographic and often socioeconomic background to predict whether they would be a flight or public safety risk if released pretrial. Because of this, the tools have been criticized by some as producing overly broad results that are based on social stereotypes and not the facts of a particular case.

Judges who have used these tools say the results they generate are helpful in informing their decisions, but they say they have steered away from relying on them entirely.

“Every case is different,” said San Francisco Superior Court Judge Michael B. McNaughton in a recent Daily Journal interview. “I make my decisions based on the people that are before me and the facts that are before me, informed by the [risk assessment tool], but not dictated by it. Not every crime is the same.”

The report recommends that risk assessment tools should produce separate scores for different criteria, such as criminal history or previous efforts to avoid court, as opposed to a composite score of all outcomes, which some tools currently produce. Courts are also advised in the report to use tools that have been validated among a similar population to the people in their jurisdiction to avoid possible skewed results based on stereotypes of other demographics.

The report asserted, “It may be possible to mitigate bias in pretrial practices by examining alternate proxies for new crime that are less susceptible to racial bias, such as new convictions instead of new arrest, or by focusing on types of arrests that are known to have less disproportionate impact across groups.”

However, it did not produce any recommendations, citing the need for further research.