Prisons Chief Pledges to Implement Criminal-Justice Reform

WASHINGTON (CN) – The head of the Bureau of Prisons on Thursday reaffirmed the Trump administration’s commitment to implementing the landmark criminal-justice bill signed into law last year, even as reform advocates raised concerns about how some of the law’s key provisions are taking shape.

Inmates stand in the yard at Arizona State Prison-Kingman in Golden Valley, Ariz., on July 4, 2015. (Patrick Breen/The Arizona Republic via AP, File)

“We continue to remain focused on the full and balanced implementation of the First Step Act,” Bureau of Prisons Director Kathleen Hawk Sawyer told lawmakers Thursday afternoon at a hearing before the House Subcommittee on Crime, Terrorism and Homeland Security.

The First Step Act became law in 2018, and was among the last actions Republicans took while they still held unified control of government in Washington. The bill was a bipartisan criminal-justice reform package years in the making and reflected a consensus that emerged between Republicans and Democrats on the subject.

A group of 10 people who were released from prison early thanks to the First Step Act’s reforms gathered in the hearing room to listen on as lawmakers received updates on how the law is being implemented.

In addition to scaling back mandatory minimum sentences for certain crimes and making retroactive a law that reduced the sentencing disparity between crack and powdered cocaine, the bill required the creation of a tool to evaluate the risk that an individual inmate will end up back in prison after being released.

The tool, the completion of which the Justice Department announced in July, will help pair inmates with programs aimed at reducing their risk of recidivism and that can let them earn credit towards a reduced sentence.

Associate Deputy Attorney General Antoinette Bacon told lawmakers Thursday that the Justice Department is in the process of evaluating comments it received on an initial version of the tool, known as the Prisoner Assessment Tool Targeting Estimated Risk and Needs, or PATTERN. Bacon said the agency hopes to be able to publish the updated tool in the near future.

The tool is at the heart of the First Step Act’s prison reforms, but also of advocates’ concerns about how the Trump administration is implementing the law.

“It is vital that that scoring system and the development of it be transparent, that it be fair, that it be valid and that it be unbiased,” David Patton, executive director of Federal Defenders of New York, said at the hearing Thursday. “And I have to say, unfortunately, I have some serious concerns about those at the outset.”

To create the tool, the National Institute of Justice contracted with two academics, who worked with the NIJ, the Bureau of Prisons and an independent review commission the Justice Department created as required by the First Step Act.

The resulting tool looks at factors inmates cannot change while in prison, such as their age when they were first arrested, and others that they can, such as the number of drug treatment or education programs they complete while behind bars.

The tool is meant to be an improvement on the existing system the government uses to evaluate recidivism, but Melissa Hamilton, a reader of Law and Criminal Justice at the University of Surrey School of Law, said Thursday her preliminary analysis of the tool raises concerns that it returns a high rate of false positives – wrongly identifying people as more likely to land back in prison about 32% of the time. She also said she is concerned the tool has a racial bias.

John Walters, a former official in the George W. Bush administration who sits on the independent review commission, tried to assuage concerns about the tool, saying in particular the commission is working on “somewhat substantial” changes aimed in particular at removing possible bias in the tool.

He said the contractors who developed the PATTERN system have run 200 additional hours of analysis at the independent review commission’s direction with the hopes of weeding out bias.

“Obviously we want the instrument to be valid, but we also want the instrument to capture real differences and not bias,” Walters said.

%d bloggers like this: