Hey Commissioner... Hand Me That Predictive Tool (Part 2)
Can the tools of psychology prevent crime? (2 of 2)
Posted Jun 01, 2009
When I tutored Federal Penitentiary inmates decades ago, I made a big mistake. Part Two of "Hey Commissioner: Hand me that Predictive Tool" is about how mistakes like this happen, and what they mean. In Part One of this two-part post, I discussed the virtues of predictive tools when deciding which prisoners to parole. But people abandon the available predictive instruments and instead try to peer into the prisoner's soul, claiming their intuitive judgments are more accurate.
How do parole boards decide instead? Most have the demographic information in the file, as well as a record of the inmate's prior behavior. But human nature being what it is, some of this valuable diagnostic information is simply ignored, and all of it is improperly weighted. In a National Public Radio interview, a recent Commissioner of the Maryland Board of Parole explained his own method for making parole decisions: "You look in their eyes; you can feel, you know, if they're being sincere or not. And you learn to sort of see right through them."
Apparently their souls are not transparent, but there is a reason I don't laugh too hard at this conceit. In 1981, a parole board convened to hear the case of an inmate I was tutoring at the Lewisburg Federal Penitentiary. The inmate's case worker suggested I speak on his behalf at the parole hearing. My experiences with this inmate were good; he worked hard on his assignments and was on track to take the GED. My last testament was for an inmate who, I later discovered, ran a vicious rape ring while in prison - a side of him he didn't share as we practiced our square roots. It is a special Federal Penitentiary inmate who scares other prisoners.
What went wrong when I provided testimony? First, my evidence was limited. I only saw him once a week, in the protection of the prison library. Second, the limited evidence I did have was, as a matter of fact, not diagnostic of reform. Learning to multiply fractions, acting politely, showing up promptly, etc., doesn't predict success in parole. Third, he didn't seem all that different from people I grew up with and worked with. (When I drove a truck, I was partnered with a guy who kept a gun in his boot, and would take it out on a long night drive to clean it, a bit nervously I thought.) The problem is, these experiences mattered to me. This familiarity influenced my judgment of the inmate's dangerousness. Finally, I trusted my own judgment.
The American public pays dearly when our broken parole system uses coffee clutch methods to solve the problems of a 21st century megademocracy. When parolees commit the same crime again people are harmed. When they otherwise violate their release conditions, their re-commitment is costly. In the meantime, there are many parole-eligible inmates who would not recidivate if released. And yet, our corrections system continues to employ parole hearing techniques guaranteed to err. Those methods are more suited to lazy conversation than scientific identification. As a result of this neglect, the corrections system doles out about $18,000 per year on each person denied parole by a board who would have integrated nicely, and on their own dime. This picture of the corrections system leaves out the important and sensitive issue of identifying candidate parolees. A 2002 Bureau of Justice Statistics study of recidivism indicates that the rates are highest for motor-vehicle theft (78.8%, $4000 per instance), possession or sale of stolen property (77.4%, $8000), larceny (74.6%, $370), and burglary (74.0%, $1400). Other errant parole releases can be even more costly. Basing harm costs on civil awards, a recent study, "Measuring the Costs and Benefits of Crime and Justice" placed the average cost of an instance of arson (without a fatality) at $38,000 and physical abuse of a child at $67,000.
People who build these models often wonder why people can't swallow their pride and step aside. But our arrogance rests on an unwarranted self-esteem that blinds us to our frailties. We can't shake the idea that these forecasting customs, with their bloodless scores, can't track the subtle tissue and turns of human behavior. From a psychological point of view, this impression is understandable. Inmates want to be heard. So do their victims. The courts have become increasingly sensitive to the restorative value of this closure. Inmates want people to know that there is redemption, that they believe they are not the same person who has committed the crime for which they were incarcerated. Their experience as a prisoner has transformed them, making them remorseful, more empathic, more determined to contribute to society, and more committed than ever to making a positive impact on the lives they touch. Victims and their families want others to know that their lives have been permanently altered by the acts of the person before them. It seems unfair to quiet their voice just so we can hear the data speak. But that's not really what we are doing. The alternative is patronizing. If you think that applying numbers to a human is essentially dehumanizing, talk to new parents about the Apgar score, applied to newborns within seconds of birth. This simple number guides the infant's treatment on the spot, and as it turns out, predicts complications years down the line. Our most precious gifts, embraced by numbers.
After 30 years of development, there is nothing high-risk or experimental about these parole decision-making techniques. Corrections institutions can now use commercially available software to look out years beyond an arrest or admittance, to determine whether a patient or prisoner will be violent (for just two examples, see here or here).
I honestly don't know whether my GED student, the U.S.'s inmate, was ever released on parole. In 1981, the parole board could not have known to ignore my prediction. The question is, why are parole boards still offering theirs?
J.D. Trout is a professor of philosophy at Loyola University in Chicago, and his book, The Empathy Gap: Building Bridges to the Good Life and the Good Society, recently appeared with Viking/Penguin.