Real-Time Automated Parole Risk Assessment: Confirming the Importance of Dynamic Predictors of Risk

Sharon Johnson, Applied Research Services, Inc.

ABSTRACT
The authors present their third-year results of recidivism research conducted with over 15,000 offenders who completed parole supervision in Georgia between 1999 and 2001. Existing comprehensive state-operated prison classification and parole operations databases were linked to develop two assessment instruments to aid parole officers in making supervision-level assignment decisions. These unique instruments are computer-generated - staff time devoted to completing "pencil and paper" assessments is eliminated. Computer programs executed by the Parole Board utilize real-time operations data to compute a probability score for each parolee on the first day of supervision and at regular intervals thereafter. The "estimated risk level" is currently made available to parole officers through the existing automated case management system in a statewide pilot test phase. The project is the first in Georgia to demonstrate empirically the important role dynamic risk factors play in determining parole supervision success. The authors will discuss how changes in parole activity (employment and program participation) directly influence risk for re-offending.



(Return to Program Resources)

Updated 05/20/2006