Risk Assessment Instruments

Risk assessment instruments are based on many research studies which followed released sex offenders and identified factors associated with those who re-offended. The factors are statistically weighted. The Static-99R is the most widely used such instrument. Many research studies have proven its predictive accuracy. The sexual re-offense rate for the typical sex offender is between 4% and 12% after 5 years from release from custody, and between 6-22% after 10 years. (Hanson, et al., Absolute Recidivism Rates Predicted by Static-99R and Static-2002R Sex Offender Risk Assessment Tools Vary Across Samples: A Meta-Analysis, Criminal Justice and Behavior (2012) 39: 1148.) Evaluations conducted by mental health professionals who base their opinions on interviews and reviewing criminal histories have not proven to be as accurate as using structured clinical judgment that incorporates validated risk assessment instruments.

Recent research has shown that the predictive accuracy of re-offense can be increased slightly when dynamic (changeable) factors are combined with static (unchangeable) factors. These include things like substance abuse, personality disorders, deviant sexual interests, emotional identification with children, and self-regulation problems. A sex offender in a mandated treatment program will be assessed on other risk factors by a certified treatment provider using dynamic and violence risk assessment instruments designated by the SARATSO Committee. The combination of the three instruments, static, dynamic, and violence will give a better picture of the overall risk of individuals who have previously sexually offended.


The Static-99R is based on static (unchanging) risk factors which predict the potential for sexual re-offending. This risk assessment instrument is required by law to be used by the California Department of Corrections and Rehabilitation to assess every eligible sex offender prior to release on parole; by Probation, to assess every eligible sex offender pre-sentencing and on a probation case load; and by the Department of State Hospitals, prior to release of an eligible sex offender from a DSH institution. In 2012 a study showed that inter-rater reliability on the instrument is strong in California (scorers are consistent in scoring the instrument). For more information on the Static-99R, visit www.static99.org.

Validation Studies in California Support Good Predictive Accuracy of Static-99R
The California Department of Justice (DOJ) partnered with the SARATSO Committee to conduct two validation studies of the Static-99 risk assessment instrument.

DOJ provided data on registered sex offenders in California. The data was analyzed by experts in the field of risk assessment retained by the SARATSO Committee. The 2014 study examined re-offense rates of parolee sex offenders five years after release from prison to the community. The 2016 study analyzed the sexual re-offense rates of both parolee and probation offenders five years after release from custody into the community. Both studies showed good predictive accuracy of the Static-99R in predicting risk of sexual recidivism among a diverse California sex offender population. The 2014 study is described in the Journal of Threat Assessment and Management (2014), Vol. 1, No. 2, at pp. 102-117. If funding is available to continue this study, the study will examine 10-year re-offense rates in 2017. The Static-99R was found to be very accurate in predicting who would reoffend in California, accurately predicting who would commit a new sex offense in about 82% of cases in the 2014 study. High risk offenders had a recidivism rate of over 29%, while low risk offenders had a recidivism rate of only 1.6%.

In the 2016 study (Lee, S., Hanson, R. Karl, et al., The Predictive Validity of Static-99R for Sexual Offenders in California: 2016 Update), the criminal histories of 1,626 offenders were examined to determine recidivism rates. The study concluded that overall the Static-99R works well in predicting risk of sexual re-offense among various ethnic groups. The 2016 study showed that the rate of re-offense in California was slightly lower than the average rate of re-offense found in international samples. The 2016 study also found that transient offenders reoffended sexually at a higher rate than non-transient offenders. The study concluded that transient status among both probation and parolee offenders seems to be associated with higher sexual recidivism rates.

JSORRAT-II (Juvenile Risk)

The Juvenile Sexual Offense Recidivism Risk Assessment Tool – II (JSORRAT-II) is a 12-item, empirically derived actuarial risk assessment tool for juvenile boys ages 12 to 18. It was developed to provide empirically-based estimates of risk for future juvenile sexual offending by male juveniles in the juvenile justice system for prior sexual offenses. The JSORRAT-II was based on an exhaustive sample of 636 boys adjudicated for a sexual offense and validated on a sample of 538 boys in Utah and 529 boys in Iowa adjudicated for sexual offenses. It recognizes the potential for accurate risk assessment to inform a range of decisions, including placement, programming, supervision, and other resource allocation decisions. The combined juvenile sexual recidivism rate in the two validation samples (N = 1067) was 9.65%. Based on the patterns of JSORRAT-II scores and recidivism rates, the following risk levels emerged as optimal.

Risk Level Score Range # Recidivists/# Scoring in Range Recidivism Rate Relative Risk Ratio*
1 0 1/106 0.9% 0.10
2 1-3 38/552 6.9% 0.71
3 4-7 43/317 13.6% 1.40
4 8+ 21/92 22.8% 2.36

*Relative Risk Ratio = Recidivism Rate for Risk Level/9.65% base rate

As noted in the table, the highest risk level was associated with a 22.8% juvenile sexual recidivism rate, which was 2.36 times greater that the base rate for the entire sample (9.65%). A score this high was relatively rare, occurring for only about 8.6% of the 1,067 boys in the sample.

The JSORRAT-II was developed and validated on boys adjudicated for sexual offenses in the 1990’s and early 2000’s. Since that time, base rates for juvenile sexual recidivism have steadily decreased. Caldwell (2016) examined juvenile sexual recidivism rates in 33,783 cases from 106 studies dating from 1938 to 2015. He found that the year of the study significantly predicted juvenile sexual recidivism rates, with newer studies being associated with lower base rates of juvenile sexual recidivism. The weighted mean juvenile sexual recidivism rate for studies conducted between 2000 and 2015 was 2.75%, substantially lower than the 10.3% rate for studies conducted between 1980 and 1995.

This suggests that the recidivism rates for each risk level in the table above are over-estimates of what recidivism rates would be today. If we assume that the correct base rate for juvenile sexual recidivism in contemporary society is 3%, a boy scoring 8 or higher would have an estimated recidivism rate of 3% multiplied by 2.36, which equals 7.08%. This level of risk certainly does not meet the requirement of being “well above average risk,” as required for mandatory lifetime registration in California pursuant to Senate Bill 384 (2017 Leg. Sess.)

STABLE-2007/ACUTE-2007 (Dynamic Risk)

The Stable-2007/Acute-2007 is scored by certified treatment providers working with sex offenders on probation or parole. (Pen. Code, sec. 290.09.) These tools measure dynamic (changing) risk factors which are empirically related to the risk of re-offense, and are evidence-based risk assessment tools. Dynamic risk assessment supplements the static risk assessment now done in California using the Static-99R, and gives a better picture of the overall risk of re-offense presented by sex offenders on supervision. The STABLE is predictive of the risk of future sexual offending. (See Literature Review/Validity Study of the STABLE-2007/ACUTE-2007.)

LS/CMI (Violence Risk)

The tool is predictive of the risk of violent re-offending (see Literature Review of LS/CMI), and is an evidence-based risk assessment tool. The LS/CMI is used by sex offender management professionals to assess registered sex offenders while they are on probation or parole. (Pen. Code, § 290.09.) Violence risk assessment supplements the static and dynamic risk assessments.