1. November 5, 2014 New Nonviolent Offender Risk Assessment Instruments – Status Update VIRGINIA CRIMINAL SENTENCING COMMISSION

2. In 1994, Virginia’s legislature directed the newly- created Sentencing Commission to: Develop an empirically-based risk assessment instrument predictive of a felon’s relative risk to public safety, and Apply the instrument to nonviolent felons recommended for prison, with a goal of placing 25% of those offenders in alternative sanctions. After the instrument was developed and pilot- tested, the Commission recommended, and the General Assembly approved, statewide implementation for July 1, 2002. Legislative Directive for Nonviolent Offender Risk Assessment 2

3. In 2003, the General Assembly directed the Commission to determine, with due regard for public safety, the feasibility of adjusting the instrument threshold to recommend additional low-risk nonviolent offenders for alternative punishment. The Commission concluded that the threshold could be raised from 35 to 38 points without significant risk to public safety. − Change became effective July 1, 2004. Legislative Directive to Revisit Nonviolent Offender Risk Assessment Legislative Directive for Nonviolent Offender Risk Assessment 3

4. 4

5. The risk assessment is completed in larceny, fraud and drug cases for offenders who are recommended for incarceration by the sentencing guidelines. Offenders recommended for probation without incarceration do not undergo risk assessment. − Goal is to avoid net widening. Offenders must also meet the eligibility criteria. − Most importantly, offenders with a current or prior violent felony conviction (as defined in § 17.1-805) are excluded from risk assessment. Use of Nonviolent Offender Risk Assessment 5

6. For offenders who score low enough on the risk scale, the sentencing guidelines cover sheet indicates a dual recommendation. − Traditional incarceration − Alternative punishment As with the sentencing guidelines, compliance with the risk assessment recommendation is discretionary. If a judge follows either sentencing recommendation, he or she is considered in compliance with the guidelines. Use of Nonviolent Offender Risk Assessment 6

7. Because it had been a number of years since the instrument was last examined, the Commission, in 2010, directed staff to begin the process of re- validating its risk assessment tool based on a more recent sample of felony cases. Following completion of the study, the Commission approved the revised risk assessment tools (one for fraud/larceny offenders and one for drug offenders). The 2012 Annual Report included a recommendation to implement the revised risk instruments, which was accepted by the General Assembly. The new instruments became effective on July 1, 2013. Most Recent Risk Assessment Study 7

8. As with prior nonviolent offender risk assessment studies, the official measure of recidivism was a new felony conviction within 3 years. However, multiple measures of recidivism were collected. Any new arrest New felony arrest Any new conviction New felony conviction New conviction is measured as a new arrest within three years of release that ultimately resulted in a conviction. Recidivism Measures 8

9. Three-Year Recidivism Rates (following release to community) Analysis is based on the sample weighted to reflect the population of offenders eligible for risk assessment. New conviction is measured as a new arrest within three years of release that ultimately resulted in a conviction. 9

10. Three-Year Recidivism Rate: New Felony Conviction within Three Years by Offense Group 10 Total = 1,509 Analysis is based on the sample weighted to reflect the population of offenders eligible for risk assessment. New conviction is measured as a new arrest within three years of release that ultimately resulted in a conviction. 10

11. Guidelines users reported that information for two factors on the previous risk assessment instrument (employment history and marital status) was not always available. It has always been the Commission's policy that the guidelines preparer err on behalf of the defendant if a particular piece of information is unknown. As a result, some offenders recommended for an alternative sanction would not have been recommended had unemployment and marital status been known and scored accurately. Scoring of Employment Record and Marital Status on FY 2003 -FY 2013 Risk Assessment Instrument 11

12. A check box was added to the Nonviolent Offender Risk Assessment form in FY2011 to indicate when offender information was not available. In addition, data entry procedures were modified to track instances where scores were missing. In FY2011, the check box was marked or information was missing relating to unemployment or marital status in 14.4% of eligible cases. Scoring of Employment Record and Marital Status on FY 2003 -FY 2013 Risk Assessment Instrument 12

13. Total = 1,509 Based on Previous Risk Assessment Instrument (as scored) Analysis is based on the sample weighted to reflect the population of offenders eligible for risk assessment. New conviction is measured as a new arrest within three years of release that ultimately resulted in a conviction. Three-Year Recidivism Rate: New Felony Conviction within Three Years 13

14. Total = 963 Based on Previous Risk Assessment Instrument For Offenders Who Received Points on the Employment or Marital Factor Three-Year Recidivism Rate: New Felony Conviction within Three Years Analysis is based on the sample weighted to reflect the population of offenders eligible for risk assessment. New conviction is measured as a new arrest within three years of release that ultimately resulted in a conviction. 14

15. MOST RECENT RISK ASSESSMENT STUDY For Drug offenders: The FY2003-FY2013 instrument correctly identified 82.6% of non-recidivists. The FY2014 instrument correctly identified 84.0% of non-recidivists. This figure was found to be stable across 750 subsamples. The threshold for the FY2014 instrument was set such that approximately the same percentage of drug offenders would be recommended for alternatives as had been in FY2012 (61.3%).

16. MOST RECENT RISK ASSESSMENT STUDY For Larceny/Fraud offenders: The FY2003-FY2013 instrument correctly identified 76.3% of non-recidivists. The FY2014 instrument correctly identified 79.3% of non-recidivists. This figure was found to be stable across 650 subsamples. The threshold for the FY2014 instrument was set such that approximately the same percentage of larceny/fraud offenders would be recommended for alternatives as had been in FY2012 (42.6%).

17. Recommended for Alternative Not Recommended for Alternative N=6,062 N=7,060 N=6,704 Risk Assessment Outcomes for Nonviolent Offenders* * Offenders recommended by the sentencing guidelines for prison or jail incarceration 17 N=6,200 17

18. Recommended for Alternative Not Recommended for Alternative 18 FRAUD Recommended for Alternative Not Recommended for Alternative LARCENY Recommended for Alternative Not Recommended for Alternative SCHEDULE I/II DRUG Recommended for Alternative Not Recommended for Alternative OTHER DRUG Risk Assessment Outcomes for Nonviolent Offenders* * Offenders recommended by the sentencing guidelines for prison or jail incarceration N=1,025 N=806 N=2,537 N=2,436 N=2,549 N=2,400 N=594 N=560 18

19. 19 FY2005-FY2006*FY2014 Fraud55.7% Male62.9% Male Fraud 65.9% Legally Restrained at Offense 68.3% Legally Restrained at Offense Larceny31.4% Female37.6% Female Other Drug 5.0 median prior record points scored on Section C 7.0 median prior record points scored on Section C Examples of Differences in Offender Population: FY2005-FY2006 versus FY2014 * Population used to create sample for the most recent risk assessment study. 19

20. While the analysis to develop the new risk assessment instruments was valid, the proportion of certain types of individuals in the felon population appears to have changed. These shifts in the population appear to be affecting risk assessment outcomes (i.e., the percentage of offenders recommended for alternatives). Staff will continue to monitor the new risk assessment instruments throughout FY2015. Monitoring 20