1. Use of Offender Risk Assessment in Virginia Presentation at the 2012 NASC Conference Meredith Farrar-Owens Virginia Criminal Sentencing Commission

2. Risk Assessment in Virginia In Virginia, risk assessment has become an increasingly formal process Nonviolent offender risk assessment Sex offender risk assessment The goal was to produce an instrument that is broadly accurate and provides additional useful information to judges Risk assessment is integrated into the sentencing guidelines and is designed to avoid net widening 2

3. Abolish parole Establish truth-in-sentencing Convicted felons must serve at least 85% of the pronounced sentence Target violent felons for longer terms of incarceration/incapacitation Expand alternative punishment options for some nonviolent felons in order ensure sufficient prison capacity for violent offenders Goals of Virginia’s Sentencing Reform Legislation (1994) 3

4. 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 to determine appropriate candidates for alternative sanctions Apply the instrument to nonviolent felons recommended for prison, with a goal of placing 25% of those offenders in alternative sanctions Legislative Directive for Nonviolent Offender Risk Assessment 4

5. The Commission studied 1,500 property and drug felons and examined over 200 factors relating to criminal record, substance abuse, education, employment, etc. Recidivism was defined as a new felony conviction within three years A risk assessment worksheet was developed based on the factors that were statistically relevant in predicting recidivism The instrument was pilot tested in six circuits during 1997-2001 Development of Nonviolent Offender Risk Assessment 5

6. Offender Reconviction Rates and Cumulative Proportion of Affected Offenders Recommended for Alternative Punishment Cumulative Proportion of Affected Offenders Risk Assessment Score 0-23-45-78-910-1112-1415-1718 & up 0% 20% 40% 60% 80% 100% 25% 12% Offender Reconviction Rate 6

7. The NCSC conducted an independent evaluation of the risk assessment instrument used in the pilot sites Evaluators concluded: “Virginia's risk assessment instrument provides an objective, reliable, transparent, and more accurate alternative to assessing an offender’s potential for recidivism than the traditional reliance on judicial intuition or perceptual short hand” Independent Evaluation by the National Center for State Courts ( NCSC ) 7

8. Cost-benefit analysis suggested a net benefit Evaluators recommended that the instrument be refined based on more recent cases and then expanded statewide Independent Evaluation by the National Center for State Courts ( NCSC ) Savings Reduced use of prison and jail $8.7 million Costs Alternative sanctions $6.2 million Costs Offender recidivism $1 million Net benefit in pilot sites $1.5 million – – = 8

9. The Commission updated the risk assessment instrument, testing and refining the scale using more recent felony cases The Commission recommended, and the legislature approved, that the risk assessment be expanded statewide Statewide implementation began July 1, 2002 Refinement of Nonviolent Offender Risk Assessment 9

10. Significant Factors in Assessing Risk for Nonviolent Offenders Never Married by Age 26 Additional Offenses Prior Arrest w / in Past 18 Mos. Prior Adult Incarcerations Male Offender Not Regularly Employed Offense Type Prior Felony Record (Adult and Juv.) Offender Age Relative Degree of Importance 10

11. The risk assessment is completed in larceny, fraud and drug cases for offenders who are recommended for incarceration by the sentencing guidelines who also meet the eligibility criteria − Excludes offenders with a current or prior violent felony conviction − Excludes offenders who sell 1 oz. or more of cocaine − Excludes offenders who must serve a mandatory term of incarceration Use of Nonviolent Offender Risk Assessment 11

12. 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 12

13. Felony Drug, Fraud and Larceny Convictions Prison In/Out Decision Guidelines Section A No Prison Prison Section B Probation/Jail Decision Section C Prison Length Decision Non-incarceration Recommendation Alternative Punishment Recommendation Jail Incarceration Sentence ProbationJail Section D Risk Assessment Alternative Punishment Recommendation Prison Incarceration Sentence Section D Risk Assessment Use of Nonviolent Offender Risk Assessment 13

14. 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 Sentencing 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 14

15. New Risk Assessment Threshold Old Risk Assessment Threshold Score Percent of offenders Reconviction rate for offenders scoring at or below point value 352.5% 12.4% 362.7% 13.9% 372.2% 13.4% 382.7% 13.6% 395.4% 16.0% 403.0% 18.8% More than 4058.7% Nonviolent Offender Risk Assessment by Score and Cumulative Reconviction Rate 15

16. Recommended for Alternative Not Recommended for Alternative N=6,062 N=6,981 N=6,473 Risk Assessment Outcomes for Nonviolent Offenders* * Offenders recommended by the sentencing guidelines for prison or jail incarceration 16

17. Received an Alternative Sanction Did Not Receive an Alternative Sanction * Sentencing guidelines recommendation is for incarceration with a midpoint of one year or more Risk Assessment Outcomes for Nonviolent Offenders ( as applied to those recommended for PRISON incarceration ) * Recommended for Alternative Not Recommended for Alternative N=3,583 (1,868) (901)(967) 17

18. It assesses risk only and not needs It does not include dynamic risk factors Young, unemployed, unmarried men are much less likely to be recommended for an alternative (it takes only a short prior record to push the score over the threshold) Criticisms of Virginia’s Risk Assessment Instrument 18

19. Virginia’s risk assessment instrument is intended to be a quick, easy-to-administer tool based on information readily available at sentencing Nonviolent offender risk assessment can only help offenders The sentencing guidelines for these offenders recommend incarceration Risk assessment identifies the lowest risk of these offenders and recommends alternative punishment Response 19

20. Develop a sex offender risk assessment instrument based on the risk of re-offending and the impact of treatment interventions Integrate a risk assessment instrument into the sentencing guidelines for sex offenses Determine the range of sentences that should be imposed on convicted sex offenders Legislative Directive for Sex Offender Risk Assessment 20

21. Studied 600 felony sex offenders released from incarceration (or given probation) Offenders followed for 5-10 years after return to the community Recidivism defined as a re-arrest for a sex offense or other crime against the person Concern that reconviction drastically underestimates recidivism due to difficulties in detection/prosecution of sex offenses Development of Sex Offender Risk Assessment 21

22. Offender Age Prior Person/Sex Arrests (Felony and Misd) Offense Location Employment History Offender Relationship/Victim Age Prior Incarcerations Education No Prior Treatment Aggravated Sex. Battery with Penetration Significant Factors in Assessing Risk for Sex Offenders Relative Degree of Importance 22

23. Risk Assessment Score Recidivism Rate Rates of Recidivism by Risk Assessment Score 23

24. Offenders scoring 28 or more are always recommended for prison and the upper end of the recommended prison sentence range is increased as follows: Risk Assessment Score Recommended Range Adjustment 44 or more Increase upper end of range by 300% 34 to 43 Increase upper end of range by 100% 28 to 33 Increase upper end of range by 50% Up to 27 No change Midpoint recommendation and low end of the recommended range remain unchanged Use of Sex Offender Risk Assessment 24

25. Sentencing Guidelines Compliance Rates for Rapists by Risk Assessment Levels FY 2011 Moderate Risk High Risk Very High Risk 30% 9% 15% No Level23% 50% 55% 72% 58% 20% 30% 9% --- 0% 6% 4% 19% 10 33 47 120 Risk Assessment Level Below Guidelines Compliance Above Guidelines Number of Cases Traditional Range Adjusted Range 25

26. Challenges to Integration of Risk Assessment into the Sentencing Guidelines

27. American Civil Liberties Union (ACLU) urged Virginia legislators to block implementation of offender risk assessment (2001), stating that: Statistical correlations are not a legitimate basis for assessing criminal penalties Virginia is the first ( and maybe only ) state to base criminal sentences on generalized, actuarial data ACLU Challenge 27

28. Applying the risk assessment punishes offenders based upon “status” in violation of the cruel and unusual punishment clause Risk assessment violates the due process requirement of fundamental fairness in criminal proceedings because the sentence is based not on the offender’s record or crime, but on the characteristics of other offenders in other crimes ACLU Challenge Eighth Amendment Violation Fourteenth Amendment Violation 28

29. ACLU did not persuade Virginia’s legislators Legislature approved the statewide implementation of risk assessment for nonviolent offenders beginning July 2002 ACLU Challenge 29

30. Virginia’s Court of Appeals has repeatedly refused to interfere with judicial reference to offender risk assessments “The discretionary sentencing guidelines are not binding on the trial judge; rather, the guidelines are merely a tool to assist the judge in fixing an appropriate punishment” “When a sentence falls within the statutory limits set by the legislature, this court will not interfere with the judgment” Virginia Court of Appeals (2004) Court Challenges 30

31. Meredith Farrar-Owens Virginia Criminal Sentencing Commission meredith.farrar-owens@vcsc.virginia.gov