"‘Most folks make the assumption that parole boards are no longer on the scene,” says Edward Rhine, director of the parole release and revocation project at the University of Minnesota. Though this is true in the federal corrections system, almost all states still have parole boards. And most of those boards still wield a great deal of control over who stays in prison, who’s let out and who gets sent back for violating the terms of his release.

As you’ve likely read in the newspaper or heard on your nightly news station, these boards don’t always make the best decisions. Unfortunately, although efforts have been made to improve them, there is still a long way to go. In more than half of the states, one obstacle is the lack of adequate qualifications for the men and women who are selected -- generally by governors -- for these positions. You might think that there would be a high bar for people who hold the keys to prison cells in their hands. In fact, out of 45 states that responded to a survey run by the University of Minnesota, 19 indicated that they had no mandated qualifications at all. In other words, a high school dropout could conceivably be appointed. Of the 25 states that had mandated qualifications, 10 required a college degree but no pertinent experience. Only 14 required some years of experience in criminal justice or aligned fields of endeavor.

On the more positive side, risk assessment tools are making their way into parole systems, incorporating data to analyze the likelihood of recidivism. The data “gives the parole board more support to estimate the potential [of an individual] to commit a crime,” says Jeff Anderson, manager of the Office of Research and Planning at the Michigan Department of Corrections. Among the factors used in various states’ risk assessments are whether the inmate was convicted of a violent crime; the age of first offense; the number of violent and nonviolent crimes committed in the past; educational levels; history of criminal acts in the potential parolee’s family; and employment history.

This is critical information. The No. 1 goal when an inmate is released is that he or she pose no danger to society.

Michigan began to change its parole approach almost a decade ago. It had become clear that heinous crimes had led Michigan’s parole board to be acutely risk averse in its work and that was one of the reasons the prison population was escalating. In 2007, the Michigan Prisoner Reentry Initiative was put in place statewide. This program, while heavily reliant on predictive data, also focuses on the role of education in preparing a prisoner for parole. Often, states have experienced difficulties when they release prisoners who have not acquired sufficient life skills to function well in society. Without those skills, parolees are more likely to fall into old habits, including drug use and crime.

In studying the data pulled together by the initiative, Michigan discovered other important findings. For example, the corrections department learned that it was unwise to release inmates right before the weekend: Weekends tend to be party times, with little opportunity to seek employment and more opportunities to get into trouble.

Pennsylvania has been a national leader in the use of risk assessment information to guide parole decisions. Bret Bucklen, director of planning research and statistics for the Pennsylvania Department of Corrections, notes that he has a staff of 10 researchers to work with the data and that the parole department has seven. “Most states,” he says, “are lucky if they have one or two.” 

Another way in which Pennsylvania is a standout is its growing use of non-prison facilities for parolees. These are somewhat less expensive resources for men and women who don’t need prison-quality security, but aren’t ready for life on the streets. Every contract for one of these facilities is based on the performance of those placed in its charge. According to Bucklen, a facility is evaluated based on recidivism rates. If the rate falls below 15 percent, the facility earns a bonus. If the rate climbs above 20 percent, it is put on a warning status. The second time, the state cancels the contract. Contracts are reviewed every six months.

While Pennsylvania and other states are relying ever more on sophisticated risk assessments, the mere existence of these tools doesn’t mean that a parole board necessarily has to follow their guidance. The lure to seem tough on crime can be very potent. According to a New York Times article this past June, the New York State Parole Board “routinely ignores state laws and regulations that require it to consider an inmate’s rehabilitation, and to use a sophisticated risk assessment tool. Because the board members have broad discretion it has been difficult to hold them accountable.” 

Even when states do rely heavily on algorithms to help select good candidates for parole, that information isn’t always passed on to prisoners. In Utah, for example, so-called rationale sheets were used to tell inmates whether the data led to a recommendation to release them or not, but nothing more. “It was minimal in information,” says Kade Minchey, audit manager at the Office of the Legislative Auditor General in Utah. 

The opacity of this information prevented the prisoners who were turned down for parole from understanding behavioral patterns that might have led to this outcome. It didn’t help them to recognize the ways in which better behavior might help them achieve parole -- and succeed in life. Utah is currently working on providing more information to its potential parolees.

The idea behind these management trends is not to make it easier for prisoners to win parole but to help parole officials be more precise and informed about the decisions they make. 

*In earlier version of this column, Bret Bucklen's name was misspelled.