Tools of the Trade: Risk Prediction Instruments

1. Tools of the Trade: Risk Prediction Instruments Professor James Byrne University of Massachusetts, Lowell Course: Technology and the Criminal Justice System September 24, 2015

2. Risk Assessment and Crime prevention • In today’s class, we will examine the specific “tools of the trade” used by ‘experts” to predict violence in a wide range of contexts: • (1) in the general community, • (2) in schools, • (3) in the workplace, • (4) in courtrooms, • (5) in correctional facilities, and • (6) while on probation or parole. • We will begin by watching a brief lecture on the use of risk assessment to support sentencing decisions by Stanford University Professor Joan Petersilia: https://www.youtube.com/watch?v=YjPm-gedWo8

3. Presentation Overview: Risky Business is Big Business • Prediction Basics: • false positives vs. false negatives • Actuarial vs. Clinical prediction • Predicting violence among known offender populations: the needle in the haystack problem • Risk vs. Stakes: Issues to consider at key decision points: • Prevention Decisions • Apprehension Decisions • Pre-trial Release Decisions • Sentencing Decisions • Release from Custody and Community Supervision Decisions • (5) Global Variations in Pre-trial Detention and incarceration Rates • (6) Types of Risk Assessment Instruments: Private Sector Proprietary Instruments vs. Free ware

4. The False Positives Problem • (1) What do researchers mean when they talk about false positives and false negatives? Why does it matter? • False Positives are individuals predicted to be violent but who are not. • False Negatives are individuals predicted to be non-violent who turn out to be violent • Which mistake are YOU willing to make?

5. Actuarial vs. Clinical Prediction • (2) What is the difference between actuarial and clinical prediction methods? • Actuarial instruments attach specific statistical weighting to different variables which assess the risk. • They are premised on the idea that, if accuracy of prediction is the most important factor, it is best to find out how members of a comparable group of individuals conducted themselves over time.

6. Clinical Assessments of Risk • Structured Clinical Guides, in contrast, invite clinicians to consider a number of variables which will have some application to the assessment of risk in the case under consideration. • This type of assessment is based on the idea that a great deal has been learned over the past two decades about the factors which should be taken in account when conducting risk assessments on various types of mental health, forensic, and correctional populations.

7. Actuarial Vs. Clinical Risk Assessment • Which type of assessment is more accurate? • There is a large body of research that demonstrates that actuarial risk assessment instruments outperform clinical risk assessment instruments. Note: See the article on this topic in our course website’s materials section. • Grove, Zale, Lebow, Snitz, and Nelson (2000) reported that statistical • prediction was about 10% more accurate than clinical prediction and was • consistently superior across date and source of publication, type of judge • (medical vs. psychological), general or task-relevant experience, type of • data (e.g., interview results, psychological tests, trait ratings, behavioral • observations, criminal record), and amount of data available. • Statistical methods predicted forensic outcomes (e.g., criminal and violent behavior) especially well, mean effect size (d) of .89, but also fared well when predicting other outcomes.

8. Predicting Violence • (3) Once an individual is identified as a violent offender—typically because he/she has been convicted of a violent crime—can we accurately “predict” whether these individuals will commit another violent crime in their lifetime? • The short answer is no. Individual prediction is likely to be inaccurate

9. Risk Prediction for Murderers and Rapists • As a group, convicted murderers have very low recidivism rates for the same crime, but they do recidivate at moderate levels overall: • Homicide: 40.7% of homicide offenders released from prison were rearrested for a new crime (not necessarily a new homicide) within 3 years; 1.2% were re-arrested for another homicide. • Rapists have very low recidivism rates overall, but the risk posed varies by the type of sexual assault: • Rape: 46.0% of released rapists were rearrested within 3 years for some type of felony or serious misdemeanor (not necessarily another violent sex offense); 2.5% were re-arrested for another rape during this review period.

10. Risk vs. Stakes: Issues to Consider • 1. Prevention: assess risk of potential violent acts by targeted individuals at school, workplace, home, or in community—involuntary civil commitment • 2. Apprehension: individuals who abscond from probation or parole, violate restraining orders, have outstanding warrants • 3. Pre-Trial Release: risk of pretrial crime, risk of failure to appear in court assessed • 4.Sentencing: risk of likely re-offending if not incarcerated assessed • 5. Release/ Reentry: risk of recidivism upon release or at sentencing

11. Countries with the Highest Pre-Trial Detention Rates

12. Countries with the Lowest Pre-Troial detention Rates

13. Countries with High vs Low Prison Population Rates

14. A Look at “Cutting Edge” Risk Prediction Instruments • General Recidivism • Level of Service Inventory (LSI-R)General Statistical Information on Recidivism (GSIR) Youth Level of Service Inventory (YLSI)Early Assessment Risk for Boys (EARL-20B)*for childrenEarly Assessment Risk for Girls (EARL-21G) • Workplace RiskWorkplace Risk Assessment (WRA-20)Employee Risk Assessment (ERA-20) •  Spousal Violence RiskSpousal Assault Risk Assessment Guide (SARA) • Ed Latessa, Professor, University of Cincinnati, explains why we use risk assessment for our known offender population: https://www.youtube.com/watch?v=1EA3eoMvoNY

15. Risk Instruments • Violent RecidivismHare Psychopathy Checklist Revised (PCL-R) • Historical Clinical Risk -20 (HCR-20)Violent Risk Appraisal Guide (VRAG) • Hare Psychopathy Checklist, Revised - Youth Version (PCL-R: YV) • Structured Assessment of Violence Risk for Youth (SAVRY) • Sexual RecidivismSex Offender Risk Appraisal Guide (SORAG)Sexual Violence Risk-20 (SVR-20)Rapid Risk Assessment for Sex Offence Recidivism (RRASOR)STATIC-99/ STATIC 2002Minnesota Sex Offender Screening Tool - Revised (MnSORT-R)Sex Offender Needs Assessment Rating (SONAR)Estimate of Risk of Adolescent Sexual Offence Recidivism (ERASOR)