This posting is a follow-up to my earlier post, A New Era in Criminal Sentencing and Incarceration, which described the growing unease over the high rates of incarceration in the U.S. Prison overcrowding and the disparate impact of incarceration on certain populations and communities are topics with important ethical implications. Mandatory minimums and “three strikes” sentencing schemes have been controversial since their inception. In the face of pressure to address the large numbers of prisoners and mounting concern over the psychological, social, and financial consequences of long-term incarceration on individuals and communities, legislatures seem poised to act. However, replacing current sentencing schemes is no easy task; both individualized and standardized approaches are fraught with difficulties.
For more than a decade—since before the time when I was his student at UVA Law—John Monahan has been developing risk assessment models to predict the future behavior of offenders. This work, supported by the National Institutes of Health and MacArthur Foundation, has resulted in many articles (and a book). Monahan’s latest article, with frequent coauthor, Jennifer Skeem, is titled Risk Assessment in Criminal Sentencing. It can be found here: Risk Assessment in Criminal Sentencing
The use of demographic and other information, along with statistics to determine future behavior is quite different from the more traditional “gut intuition” approach. Monahan has noted that the two methods were described in Paul Meehl’s 1954 article, Clinical Versus Statistical Prediction (1954):
[…there are ]two ways of forecasting behavior. One, a formal method, uses an equation, a formula, a graph, or an actuarial table to arrive at a probability, or expected value, of some outcome; the other method relies on an informal, “in the head,” impressionistic, subjective conclusion, reached… by a human clinical judge.
[From John Monahan, Violence Risk Assessment: Scientific Validity and Evidentiary Admissibility, 57 Wash. & Lee L. Rev. 901 (2000): Violence Risk Assessment ]
The Federal Sentencing Guidelines (FSG) can be viewed as an attempt to strike a balance between the human decision-maker and the actuarial design. The guidelines are equitable in that they establish consistency across jurisdictions among like offenders, but they fail to take some of the more individualistic information (such as marital status and education level) into account. Although the FSG address several problems with the judicial discretion model, critics have pointed out the guidelines achieve uniformity at the expense of fairness, in that factors indicating the appropriateness of a lighter sentence are ignored.
Consideration of factors that have been identified and demonstrated to be reliable predictors of future criminal behavior may strike the perfect balance. On the one hand, this approach respects the individuality of each offender, because it takes into consideration that offender’s characteristics and background. On the other hand, statistics-based risk assessment decisions would, at least in some senses, treat similarly situated defendants similarly.
Of course, predictions of future violence have been a central part of the civil commitment process for years. Mental health professionals have been tasked with examining individuals and making assessments about future violence prediction. The advantage of a pure data-driven approach is increased accuracy and cognitive errors. In the criminal sentencing context, an actuarial approach lessens the impact of the personal preferences and biases of judge and jury. Ultimately, the argument goes, risk prediction models are likely to result in better accuracy, and therefore more “correct” outcomes.
Although some factors related to risk of reoffending are common in sentencing schemes, of late, primary or wholesale reliance on statistical prediction of future risk has not been. That may be changing. Courts in a number of jurisdictions appear to be incorporating risk prediction into sentencing decisions. At least one state is formally considering the move. Several months ago, news outlet fivethirtyeight reported that Pennsylvania would be the first state to experiment with a new sentencing practice, based upon risk assessment. According to the report:
Pennsylvania is about to take a step most states have until now resisted for adult defendants: using risk assessment in sentencing itself. A state commission is putting the finishing touches on a plan that, if implemented as expected, could allow some offenders considered low risk to get shorter prison sentences than they would otherwise or avoid incarceration entirely. Those deemed high risk could spend more time behind bars. [For more, see: Prison Reform Risk Assessment]
Risk assessment tools could help Pennsylvania effectively distinguish between the offenders who are most likely to pose a risk to society and those who would benefit from rehabilitation, substance abuse or other treatment, or a simple second chance. The new method may allow a state to address prison overcrowding in an effective and humane way, and it could result in less crime down the road. However, a bevy of ethical issues arise as well. As fivethirtyeight notes,
The risk assessment trend is controversial. Critics have raised numerous questions: Is it fair to make decisions in an individual case based on what similar offenders have done in the past? Is it acceptable to use characteristics that might be associated with race or socioeconomic status, such as the criminal record of a person’s parents? And even if states can resolve such philosophical questions, there are also practical ones: What to do about unreliable data? Which of the many available tools — some of them licensed by for-profit companies — should policymakers choose? Prison Reform Risk Assessment
As states and the federal government address the thorny issue of how to eliminate prison overcrowding and lower incarceration rates, actuarial data derived from social science will likely play a major role. Time will tell whether these new risk assessment tools can be implemented in a way that is ethical and effective.