Predicting Pretrial Success

Donna Lyons 2/1/2014


Criminal justice policy is using science to predict risk, helping courts make decisions about the conditions of pretrial release.

By Donna Lyons

Most criminal defendants have a right to be released prior to trial, with exceptions made only for those accused of the most serious and dangerous crimes. Yet many stay in jail because they can’t meet conditions of release, which often are monetary. Courts determine conditions of release based on safety and flight risk. But judges usually have little information about the risks defendants pose to guide these pretrial decisions.

That’s changing in a growing number of states, however. Lawmakers in at least 11 states now provide statewide guidance to judges on using results of a risk assessment in pretrial decision making. While considerations of risk have guided legislative policies and how courts exercise discretion in pretrial release, new scientific tools are making these determinations more empirical and reliable.

Lawmakers in Colorado last year addressed many aspects of pretrial release, including risk assessment, in legislation sponsored by then-Speaker Pro Tem Claire Levy (D). The new law requires courts to use individualized, evidence-based decision- making practices when setting bonds and other conditions of release. In Colorado and elsewhere, conditions typically include commercial or other secured bonds, cash, property and supervision. The law presumes that all those in custody for bailable offenses be released under the “least-restrictive conditions,” while allowing conditions that address specific concerns. Fifteen states have such requirements.

The Colorado law also allows those granted but unable to pay the financial bond to ask the court seven days after the bond was set to reconsider requiring the monetary condition of release. Levy said this is intended to reduce the significant number of people who haven’t been convicted of a crime but are in jail because they can’t afford bail, even though they don’t pose a risk of flight or a danger to public safety.

The Role of Research

“We’ve learned that assessment can reduce the number of low-risk people in county jails, and help courts distinguish those who pose the greatest risk,” says Anne Milgram, vice president of criminal justice for the Laura and John Arnold Foundation in New York. Milgram, who served as New Jersey’s attorney general before joining the Arnold Foundation, noted the Foundation’s research revealed that nearly half of the highest-risk defendants were obtaining release before trial. Yet, 90 percent of jurisdictions today do not use risk assessment instruments that can help guide these determinations, according to Milgram.

Over the past two years, the foundation has conducted

extensive research on pretrial decision making and has developed the Public Safety Assessment-Court (PSACourt) tool to help distinguish among defendants at different risk levels. The tool determines how likely a defendant is to fail to return to court or commit a new crime if released. It also identifies those defendants who are most likely to commit a new violent crime.

The PSA-Court tool was developed and validated by a research team that studied hundreds of thousands of pretrial cases in more than 300 U.S. jurisdictions. They identified the factors that were the best predictors of new crime, new violence and failure to appear. They determined that the risk assessment tool—without a time-consuming interview—can reliably predict risk posed by a given defendant.

Kentucky is the first pilot site for the tool. All 120 counties started using it in July 2013, and other pilot sites will begin this year, followed by a national rollout. Preliminary results in Kentucky are confirming the tool’s value in predicting pretrial success. Kentucky legislation in 2011 paved the way. It broadly addressed criminal justice reforms, and required judges to use results of an interview tool designed to measure a defendant’s risk of flight and potential for criminal conduct while on pretrial release pending case disposition. The process effectively gauged risk and improved pretrial release success rates, but it was labor intensive. The new PSA-Court tool changes what was a 20-minute interview process to a five-minute assessment, according to Tara Klute, director of Kentucky Pretrial Services.

“We’re finding it saves time and resources and produces better results,” Klute says.             Kentucky’s legislation also requires that low-risk and some moderate-risk defendants be released on their own recognizance or on unsecured bond. Since that requirement was put in place, non-financial releases have risen 15 percent. This adds up to 11,000 more defendants released since the new policy was put in place compared to the same time frame before the law, according to a recent report of the state’s Administrative Office of the Courts. Representative John Tilley (D), who sponsored the 2011 legislation, says the changes have saved Kentucky counties $30 million to $40 million in pretrial detention costs. “And in a state that already was safe, crime has fallen,” Tilley says of the reforms.

The court report said that appearance rates are slightly up and re-arrests slightly down, even with the considerable increase in defendants released from jail pending trial. The pretrial services office has been responsible for monitoring results in each county both before and after the legislative reforms.

Each year, 12 million people are booked into jails in the United States. Even though the majority are charged with nonviolent crimes, more than 60 percent of jail inmates are awaiting trial. Many of those in jail are there because they can’t afford bail. A recent Bureau of Justice Statistics report said that about nine in 10 detained felony defendants in large urban counties had a bail amount set that they were unable to meet.

Other Arnold Foundation research is shedding light on the effect of pretrial detention. Findings released in late 2013 show that defendants held for the entire pretrial period had a three or four times greater likelihood of being sentenced to prison or jail than defendants released at some point before trial, and prison or jail sentences imposed were two to three times longer. The research also found that low- and moderate-risk defendants who were detained before trial were more likely to commit new crimes within two years of case disposition. Arnold analysis in two states also found that moderate- and high-risk defendants released who received pretrial supervision were significantly more likely to appear in court as scheduled.

Traditionally, criminal justice hasn’t had enough data and analysis, says Milgram, who likes to refer to Arnold’s national efforts to bring smart, statistical analysis to issues as the “moneyball” approach.

 “You can’t solve problems you don’t know you have,” Milgram says. “That’s why we are collecting and analyzing data needed to answer critical questions about how we can make our communities safer, make our criminal justice system more just, and use our already scarce resources more wisely.”

States Where Courts Must Use the Results of Risk Assessment Tools in Making Pretrial Release Decisions

  • Colorado
  • Connecticut
  • Delaware
  • Hawaii
  • Illinois
  • Kansas
  • Kentucky
  • Maine
  • Oklahoma
  • South Carolina
  • Virginia

Donna Lyons is group director of the  NCSL Criminal Justice Program.

Additional Resources