Static-99R Training
Total Page：16
File Type：pdf, Size：1020Kb
Static-99R 2016 Coding Rules 8/17/2017 Jacob Bezanson, MSW 1 Why Assess Risk? o Promoting public safety o Routine interventions o Targeting scarce resources ◦ Officer time ◦ Treatment o Exceptional measures o Rank order individuals 2 Why Assess Risk? Risk Need Responsivity Model (RNR) ◦ Risk: Who should be targeted to receive the most resources ◦ Need: What should be targeted in treatment to have the greatest impact on recidivism. ◦ Responsivity: How should treatment be delivered Using validated assessments allows for the accurate adherence to these principals, and a common standardized vocabulary to communicate risk. 3 Static, Stable, and Acute Risk Factors Definitions Static – Non-changeable life factors that relate to risk for sexual recidivism, generally historical in nature Stable – Personality characteristics, skill deficits, and learned behaviors that relate to risk for sexual recidivism that may be changed through intervention Acute – Risk factors of short or unstable temporal duration that can change rapidly, generally as a result of environmental or intra-personal conditions 4 Three Generations of Risk Assessment (Bonta 1996) First Generation = “Clinical Judgment” ◦ Unstructured, Non-replicable, Personal Discretion ◦ Based on experience and level of knowledge of the literature ◦ Non-standard (even within same institution) ◦ Level of prediction little better than chance Second Generation = “Actuarial Assessment” ◦ Static, Actuarial, Structured, Replicable, Less open to Interpretation ◦ Based on factors empirically related to recidivism ◦ Standardized assessment, “Static” - Cannot measure change ◦ “Moderate” Levels of prediction, ROC’s upper 60’s to lower 70’s Third Generation = “Dynamic Assessment” ◦ Based on factors empirically related to recidivism ◦ Standardized assessment, Measures change ◦ Actuarial measure with dynamic factors 5 Prediction of Sexual Recidivism Measures Designed for d (95% CI) N (k) Sexual Recidivism Empirical Actuarial .67 (.63-.72) 24,089 (81) Mechanical .66 (.58-.74) 5,838 (29) Structured Judgement .46 (.29-.62) 1,131 (6) Unstructured .42 (.32-.51) 6,456 (11) 6 Evolution- Static-99R Step 1: Identify Static Risk Factors . (Hanson & Bussiere, 1996, 1998) So, what happens when you combine risk factors? . For example, how much riskier is a guy who is both young and has prior offenses? 7 Evolution- Static-99R Step 1: Identify Static Risk Factors . (Hanson & Bussiere, 1996, 1998) Step 2: Combine Risk Factors . RRASOR (Hanson, 1997) 8 Rapid Risk Assessment for Sex Offense Recidivism (RRASOR, Hanson, 1997) . Any male victim (0 = no; 1 = yes) . Any unrelated victim (0 = no; 1 = yes) . Prior sex offenses (3 points) . Age (1 = 18-24.9; 0 = 25+) 9 Evolution- Static-99R RRASOR Static- 99 SACJ-MIN 10 Evolution- Static-99R Original Coding Rules 2003 Static-99 was revised in with a new age weights for Item 1, renamed the Static-99R New Static-99R Coding Rules 2016 11 Sex Offender Risk Assessment Basis for STATIC-99R STATIC (unchangeable) factors o Age o Ever Lived With ... o Current non-sex violence o Prior non-sex violence o Prior sex offense s o 4+ sentencing dates o Non-contact sex offense(s) o Unrelated victims o Stranger victims Male victims o 12 Who Can I Use This With? Population STATIC-99R STABLE-2007 ACUTE-2007 Adult Male Sexual Offenders, with at least one “Category A” offense Adult Male Sexual Offenders with only “Category B” offense history Juvenile offenders With Caution With Caution With Caution aged 17 (released age 18 or older) Juvenile offenders 16 years and less Adult female 13 offenders Risk for Female Sexual Offenders Low Sexual Recidivism Rates ◦ Cortoni, Hanson, & Coache (2011) found recidivism rates less than 3% (10 studies, n = 2,490) Recidivism Risk Factors are Unknown ◦ We recommend that you use a risk scale designed for general or violent recidivism that has been validated on female offenders (e.g., LS/CMI, Andrews & Wormith, 2004) 14 Static-99R with Adolescents who Sexually Offend (Page 14) Can be used on offenders who are released after age 18 and their sex offense was at age 17 (previously was 16) AND The offense appears adult like (not antisocial with a sexual component)…otherwise do not use. 15 STATIC 99R and Gender Transformation (Page 16) Male to female transgender: . Static 99R can be used until the offender does not have a penis and has lived as a woman for at least 2 years Female to male transgender . Do not use the Static on this population as they are outside the sampling frame of the scale. 16 Generalizability Stability of findings across settings and samples Finding from Static-99R meta-analysis (23 samples; Helmus et al., 2012) ◦ Relative risk is stable (no significant variability) across diverse samples/settings ◦ Absolute risk varies significantly across samples 17 Ethnicity and Static-99R (Page 16) o Aboriginal, African-American offenders and African-Asian offenders score higher on Static-99R than Caucasian offenders o Latino offenders score lower on Static-99R than Caucasian offenders. o This does not mean that Static-99R predicts differently between these groups. o Note in your report that it predicts better for Caucasian offenders (Mdn=.76) than non-Caucasian offenders (Mdn=.70) Not stat sig difference 18 Sexual Recidivism Rates (%) 5 years 10 years 15 years Rapists 14 21 24 Incest offenders 6 9 13 Girl victim child molesters 9 13 16 Boy victim child molesters 23 28 35 19 Sexual Recidivism Rates (%) 5 years 10 years 15 years No prior sex offense(s) 10 15 19 Any prior sex offense(s) 25 32 37 20 STATIC-99R Sexual Reconviction Rates 1 0.8 Low 0.6 Medium-Low 0.4 Medium-High High 0.2 0 Years after release 21 Caveats for Predictive Accuracy: Conscientiousness effect? Completed Did Not All Steps Complete All Steps Static-99R AUC = .80 AUC = .68 AUC = .76 AUC = .58 Stable-2007 (ns) 22 DSP Outcomes – Sexual offenses “Overall” versus “Conscientious” Recidivism Rate Test ROC Overall 7.2% STATIC-99 .74 “Sexual” (57/793) S-99 & STABLE-07 .76 6.8% STATIC-99 .81 Conscientious (23/336) S-99 & STABLE-07 .84 23 “Conscientious” effect: Stuff to think about o Data from Dynamic Supervision Project (Hanson, Helmus, & Harris, 2015) o Completed all assessments requested: Static-99, STABLE- 2000, ACUTE-2000, and override for Static o Points out need for good training o Points out need for “management buy-in” o Findings say – “It works pretty well if you take it seriously” o Assessors have to be careful and consistent 24 The Static-99R Strengths o Empirically derived risk factors for sexual recidivism o Explicit rules for scoring the factors and getting a total risk score o Can rank offenders in terms of relative risk, robustly o Relatively objective instrument scoring o STATIC instruments are rarely if ever challenged in evidentiary hearings in court o Most commonly used actuarial risk scale for sexual offenders o Moderate predictive accuracy in 63 replications (Hanson & Morton-Bourgon, 2009) 25 The Static-99R Limitations o Moderate Predictive Accuracy (AU .69-.70) o Does not include all factors that predict sex offending o Absolute recidivism rates associated with specific risk scores vary across samples making estimates “more complex.” o Does not adequately measure general or violent recidivism (BARR-2002R) 26 General Scoring Issues o Stick to the Coding Rules . Even if they may not make sense to you or you do not agree. They are written that way for a reason to match the data sets on which the instrument was developed. You have to remember the intention of the scoring item, what was the original population. o Consider external risk factors (SRA-FV, VRS-SO, or STABLE-2007) o “Over-rides” not related to sexual recidivism or already considered will decrease predictive accuracy. 27 Key Concepts and Definitions 28 Standards of Proof in Decision Making for Coding Static-99R Beyond a Reasonable Doubt: Highest standard. It requires near certainty and is the standard necessary for criminal convictions. Clear and Convincing Evidence : Higher standard than the Balance of Probabilities, but it is not quite as high as Beyond a Reasonable Doubt. Balance of Probabilities (a.k.a., Preponderance of Evidence): Lowest standard. It is common in civil cases and basically means “more likely than not,” or at least 51% certainty. 29 Sexual Motivation vs. Economic Gain May or may not have sexual motivation Economic Gain: ◦ An offender arranges for another offender to molest a child but did not participate in the act for sexual motivation> It was economic. (Not a Sex Offense) ◦ Human trafficking / Pimping / Pandering for economic gain. Sexual Motivation: ◦ An offender wants to watch a video of the molestation for sexual arousal or he pimps a minor for economic gain and engages in sex with minor. 30 Sexual Motivation (Page 20-21) To count a charge or conviction as a sex offense there should be sexual o motivation for the offense, or it should be clearly part of the commission of the offense and/or based on self-report. o Non-Sexual Motivation . Forcible confinement- e.g. locked victim’s boyfriend in bathroom so he can sexually assault the victim, not sexual motivation . Break and Enter, burglary is not a sex offense unless it had a sexual motivation . Stalking motivated by sexual jealousy o Sexual Motivation . Offender followed a woman home and broke into her house with the intent to sexually assault her. He then stole some jewelry. Charged with Burglary, Theft, and Rape. The Burglary and Rape had sexual motivation and both count as sex offenses 31 Self Report in Static 99R (Pages 8-9) For immigrants and refugees from third world countries and old out-of-state records where confirmation is not possible you can use self report for criminal history. Self-report information in all other cases can not be used to substitute criminal record, it can be used to clarify behavior and help determine if an offense was sexually motivated.