Prisons turn to analytics software for parole decisions
- By Kevin McCaney
- Nov 01, 2013
The same type of analytics software police use to try to predict where crimes might occur or who is likely to commit them is also being used by prisons to help determine who isn’t likely to commit a crime — thereby hastening their parole.
At least 15 states, looking to cut costs on incarceration, now require their prison systems to use some form of risk assessment tool in evaluating inmates, and many of them are turning to predictive analytics software that looks for patterns based on a variety of factors, the Wall Street Journal reported.
The software programs, intended to complement rather than replace traditional parole assessments, measure factors such as inmates’ age when first convicted, education, whether they think their conviction was justified and whether they’re married, the Journal said. Some programs measure 50 to 100 factors overall, in contrast to the relative handful weighed by parole boards, many of whose members are political appointees without much or any training in criminology.
Adding software-driven assessments — which also can help determine the extent of supervision a parolee will require — appears to be having an effect. The Journal said populations in state and federal prisons fell by 1 percent in 2011 and appear to have fallen further in 2012, according to reports available. And that’s at least partly because of a drop in recidivism: the percentage of parolees going back to prison dropped from 15 percent in 2006 to 12 percent in 2011.
One such program is Compas (Correctional Offender Management Profiling for Alternative Sanctions), made by Northpointe, a criminal justice research and consulting company. Compas, used in Michigan, New York and other states, combines standard risk factors such as criminal history with other dynamic data in calculating an inmate’s probability of breaking parole, according to the company. It also allows for inmate interviews to be included in an assessment.
Northpointe acknowledges that Compas measures probabilities and shouldn’t be used exclusively. Tim Brennan, the company’s chief scientist, told the Journal that parole boards should override Compas’ conclusion 8 percent to 15 percent of the time. But risk assessment software does give officials a more complete view.
Analytics is an increasingly common tool in many facets of government, from detecting biological threats to managing power usage and reducing water use. Law enforcement and corrections are on board as well.
At the city level, Baltimore is using predictive software developed by Richard Berk, a criminology and statistics professor at the University of Pennsylvania, to help make decisions on parole and probation, according to the Denver Post. Philadelphia is using the same software to predict crimes.
Ohio applies the open-source Ohio Risk Assessment System (ORAS) throughout the legal process — during pretrial proceedings, in connection with community supervision, during prison intake and in preparation for parole or release.
Officials have found that a software approach, while not perfect, has helped them make better decisions and — perhaps more significantly — save money.
In a 2011 interview with the Penn Currant, Berk said the analytics programs are better than humans at combining known factors. “The predictors that we use are by and large what people have known about for a hundred years,” he said. “There’s no mystery.” But the programs are better at “squeezing more information out of those predictors” and at times finding “things that you don’t understand.”
“When you allow the computer to find it, you predict a lot better,” Berk said.
Kevin McCaney is a former editor of Defense Systems and GCN.