This is the property of the Daily Journal Corporation and fully protected by copyright. It is made available only to Daily Journal subscribers for personal or collaborative purposes and may not be distributed, reproduced, modified, stored or transferred without written permission. Please click "Reprint" to order presentation-ready copies to distribute to clients or use in commercial marketing materials or for permission to post on a website. and copyright (showing year of publication) at the bottom.
News

Criminal

Oct. 18, 2018

Risk assessment tools questioned as courts phase out money bail

Algorithmic risk assessment tools central to pretrial decisions about releasing criminal defendants are coming under increased scrutiny since SB 10, which phases out cash bail, was signed into law.


Attachments


Risk assessment tools questioned as courts phase out money bail
Albert W. Ramirez, general counsel for Golden State Bail Agents Association

Algorithmic risk assessment tools central to pretrial decisions about releasing criminal defendants are coming under increased scrutiny since SB 10, which phases out cash bail, was signed into law.

The Judicial Council is preparing to choose recommended tools.

In the meantime, according to information provided by the council, 49 of the state's 58 county courts use a type of pretrial risk assessment tool that provides judges with information about the risk of releasing a defendant before trial.

An examination of data obtained by the Daily Journal shows two tools dominate the pretrial landscape in the state: the Virginia Pretrial Risk Assessment Instrument developed by Luminosity Inc. and the Ohio Risk Assessment System, developed by Ohio's corrections department. Together, the two methods are used as pretrial risk assessment tools in 33 California counties.

Other superior courts, such as in Santa Clara and Riverside counties, use county-specific tools developed by agencies dedicated to pretrial services.

Some of the state's largest counties are using algorithms developed by private sector entities. San Francisco, for example, uses an algorithm developed by the New York-based Arnold Foundation, according to the foundation's website dedicated to its pretrial service Public Safety Assessment.

Carl McGrew, a member of the California Association of Pretrial Services and a former member of Santa Clara County's Office of Pretrial Services, said the customized approach his county took resulted in measurable success.

After implementing the pretrial risk assessment tool, the office went from overseeing 900 individuals released on their own reconnaissance to 1,400 individuals, but the percentage of recidivism remained stable, he said.

"Both pretrial and bail can point to horror stories," McGrew said in regard to scenarios where an algorithm gets it wrong and a defendant is released only to commit a different crime. "It boils down to: Are you going to be able to get out from custody without having to be wealthy?"

In San Francisco, the Arnold methodology has become a point of contention among civil rights attorneys and bail bonds agents who say the algorithm neglects crucial elements when determining, for example, whether or not the defendant poses a flight risk.

The state Supreme Court is reviewing a case challenging San Francisco's bail system, which currently co-exists with its pretrial risk assessment program. In re Humphrey, S247278 (Ca. Sup. Ct., filed Feb. 27, 2018).

The case reached the high court after the algorithmic tool received a scathing review from the 1st District Court of Appeal. In the opinion, Justice J. Anthony Kline wrote, "The one-page form risk assessment report submitted to the court by the pretrial services agency, which does not indicate a representative of the agency ever met with petitioner, provides no individualized explanation of its opaque risk assessment of petitioner and no information regarding the availability and potential for use of an unsecured bond."

Albert W. Ramirez, general counsel for the Golden State Bail Agents Association who filed an amicus brief for the Supreme Court case, said the claims that algorithmic methods are evidence-based is ironic considering the lack of evidence that they actually work.

"There's no evidence, other than self-serving studies, saying these things work," Ramirez said. "You would think something like SB 10 that creates such a fundamental change should be subjected to more rigorous testing."

Ashley Winstead, a spokesperson for Arnold Foundation, said it is funding research to evaluate the impact of broader pretrial reforms where its algorithm has been implemented.

"The PSA was developed through, and continues to be supported by, scientifically sound research," Winstead wrote in an email. "The work we are doing to validate and improve the algorithm through sophisticated statistical analysis -- not to mention the important work local jurisdictions are doing -- is an ongoing and deliberate process to ensure the PSA continues to perform the way it is intended."

#349803

Paula Lehman-Ewing

Daily Journal Staff Writer
paula_ewing@dailyjournal.com

For reprint rights or to order a copy of your photo:

Email Jeremy_Ellis@dailyjournal.com for prices.
Direct dial: 213-229-5424

Send a letter to the editor:

Email: letters@dailyjournal.com