Journal of Quantitative Criminology

, Volume 30, Issue 1, pp 79–96 | Cite as

Forecasts of Violence to Inform Sentencing Decisions

Original Paper



Recent legislation in Pennsylvania mandates that forecasts of "future dangerousness" be provided to judges when sentences are given. Similar requirements already exist in other jurisdictions. Research has shown that machine learning can lead to usefully accurate forecasts of criminal behavior in such setting. But there are settings in which there is insufficient IT infrastructure to support machine learning. The intent of this paper is provide a prototype procedure for making forecasts of future dangerousness that could be used to inform sentencing decisions when machine learning is not practical. We consider how classification trees can be improved so that they may provide an acceptable second choice.


We apply an version of classifications trees available in R, with some technical enhancements to improve tree stability. Our approach is illustrated with real data that could be used to inform sentencing decisions.


Modest sized trees grown from large samples can forecast well and in a stable fashion, especially if the small fraction of indecisive classifications are found and accounted for in a systematic manner. But machine learning is still to be preferred when practical.


Our enhanced version of classifications trees may well provide a viable alternative to machine learning when machine learning is beyond local IT capabilities.


Sentencing Forecasting Machine learning Classification trees 


  1. Banks S, Robbins PC, Silver E, Vesselinov R, Steadman HJ, Monahan J, Mulvey EP, Applebaum PS, Grisso RT, Roth LH (2004) A multiple-models approach to violence risk assessment among people with mental disorder. Crim Justice Behav 31:324–340CrossRefGoogle Scholar
  2. Barnes GC, Ahlman L, Gill C, Sherman LW, Kurtz E, Malvestuto R (2010) Low intensity community supervision for low-risk offenders: a randomized, controlled trial. J Exp Criminol 6:159-189CrossRefGoogle Scholar
  3. Berk RA (2008a) Forecasting methods in crime and justice. In: Hagan J, Schepple KL, Tyler TR (eds) Annual review of law and social science, vol 4. Annual Reviews, Palo Alto, pp 173–192Google Scholar
  4. Berk RA (2008b) Statistical learning from a regression perspective. Springer, New YorkGoogle Scholar
  5. Berk RA (2009) The role of race in forecasts of violent crime. Race Soc Probl 1(4):231–242CrossRefGoogle Scholar
  6. Berk RA (2011) Asymmetric loss functions for forecasting in criminal justice settings. J Quant Criminol 27:107–123CrossRefGoogle Scholar
  7. Berk RA (2012) Criminal justice forecasts of risk: a machine learning approach. Springer, New YorkCrossRefGoogle Scholar
  8. Berk RA (2013) Algorithmic criminology. Secur Inform 2(5) (forthcoming)Google Scholar
  9. Berk RA, Sorenson SB, He Y (2005) Developing a practical forecasting screener for domestic violence incidents. Eval Rev 29(4):358–382CrossRefGoogle Scholar
  10. Berk RA, Kriegler B, Baek J-H (2006) Forecasting dangerous inmate misconduct: an application of ensemble statistical procedures. J Quant Criminol 22(2):135–145CrossRefGoogle Scholar
  11. Berk RA, Sherman L, Barnes G, Kurtz E, Ahlman L (2009) Forecasting murder within a population of probationers and parolees: a high stakes application of statistical learning. J R Stat Soc Ser A 172(part I):191–211CrossRefGoogle Scholar
  12. Borden HG (1928) Factors predicting parole success. J Am Inst Crim Law Criminol 19:328–336CrossRefGoogle Scholar
  13. Breiman L (2001a) Random forests. Mach Learn 45:5–32CrossRefGoogle Scholar
  14. Breiman L (2001b) Statistical modeling: two cultures (with discussion). Stat Sci 16:199–231CrossRefGoogle Scholar
  15. Breiman L, Friedman JH, Olshen RA, Stone CJ (1984) Classification and regression trees. Wadsworth Press, MontereyGoogle Scholar
  16. Burgess EM (1928) Factors determining success or failure on parole. In: Bruce AA, Harno AJ, Burgess EW, Landesco EW (eds) The working of the indeterminate sentence law and the parole system in Illinois. State Board of Parole, Springfield, pp 205–249Google Scholar
  17. Bushway S (2011) Albany Law Rev 74(3)Google Scholar
  18. Casey PM, Warren RK, Elek JK (2011) Using offender risk and needs assessment information at sentencing: guidance from a national working group. National Center for State Courts.
  19. Chipman HA, George EI, McCulloch RE (2010) BART: Bayesian additive regression trees. Ann Appl Stat 4(1):266–298CrossRefGoogle Scholar
  20. Farrington DP, Tarling R (2003) Prediction in criminology. SUNY Press, AlbanyGoogle Scholar
  21. Feeley M, Simon J (1994) Actuarial justice: the emerging new criminal law. In: Nelken D (ed) The futures of criminology. Sage, London, pp 173–201Google Scholar
  22. Friedman JH (2002) Stochastic gradient boosting. Comput Stat Data Anal 38:367–378CrossRefGoogle Scholar
  23. Gottfredson SD, Moriarty LJ (2006) Statistical risk assessment: old problems and new applications. Crime Delinq 52(1):178–200CrossRefGoogle Scholar
  24. Harcourt BW (2007) Against prediction: profiling, policing, and punishing in an actuarial age. University of Chicago Press, ChicagoGoogle Scholar
  25. Hastie R, Dawes RM (2001) Rational choice in an uncertain world. Sage, Thousand OaksGoogle Scholar
  26. Hastie T, Tibshirani R, Friedman J (2009) The elements of statistical learning, 2nd edn. Springer, New YorkCrossRefGoogle Scholar
  27. Holmes S (2003) Bootstrapping phylogenetic trees: theory and methods. Stat Sci 18(2):241–255CrossRefGoogle Scholar
  28. Hyatt JM, Chanenson L, Bergstrom MH (2011) Reform in motion: the promise and profiles of incorporating risk assessments and cost-benefit analysis into Pennsylvania sentencing. Duquesne Law Rev 49(4):707–749Google Scholar
  29. Kleiman M, Ostrom BJ, Cheeman FL (2007) Using risk assessment to inform sentencing decisions for nonviolent offenders in Virginia. Crime Delinq 53(1):1–27CrossRefGoogle Scholar
  30. Kuhnert PM, Mengersen K (2003) Reliability measures for local nodes assessment in classification trees. J Comput Graph Stat 12(2):398–426CrossRefGoogle Scholar
  31. Messinger SL, Berk RA (1987) Dangerous people: a review of the NAS report on career criminals. Criminology 25(3):767–781CrossRefGoogle Scholar
  32. Monahan J (1981) Predicting violent behavior: an assessment of clinical techniques. Sage, Newbury ParkGoogle Scholar
  33. Pew Center of the States, Public Safety Performance Project (2011) Risk/needs assessment 101: science reveals new tools to manage offenders. The Pew Center of the States.
  34. Silver E, Chow-Martin L (2002) A multiple models approach to assessing recidivism risk: implications for judicial decision making. Crim Justice Behav 29:538–569CrossRefGoogle Scholar
  35. Skeem JL, Monahan J (2011) Current directions in violence risk assessment. Curr Dir Psychol Sci 21(1):38–42CrossRefGoogle Scholar
  36. Tóth N (2008) Handling classification uncertainty with decision trees in biomedical diagnostic systems. PhD thesis, Department of Measurement and Information Systems, Budapest University of Technology and EconomicsGoogle Scholar
  37. Turner S, Hess J, Jannetta J (2009) Development of the California Risk Assessment Instrument. Center for Evidence Based Corrections, The University of California, IrvineGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2013

Authors and Affiliations

  1. 1.Department of StatisticsUniversity of PennsylvaniaPhiladelphiaUSA
  2. 2.Department of CriminologyUniversity of PennsylvaniaPhiladelphiaUSA

Personalised recommendations