Brief risk scales for the prediction of sex offence recidivism

Brief risk scales for the prediction of sex offence recidivism PDF Version (8KB)

Research summary
Vol. 4 No. 5
September 1999

Question

How well can brief, structured risk scales predict sex offence recidivism?

Background

Sex offences are the focus of considerable public concern. Special policies have been developed in order to protect the public from high risk sex offenders, including intensive treatment programs, parole restrictions and long-term community supervision. All of these policies assume the ability to distinguish between low risk and high risk offenders.

Unstructured clinical judgement has had limited success in predicting which sex offenders will do it again. Risk scales designed for non-sexual criminal recidivism have been only weakly related to sex offence recidivism. Consequently, there was a need for a risk scale design specifically for sex offenders.

Method

The study examined the predictive accuracy of a brief, actuarial risk scale entitled Static-99. Static-99 was constructed from two previously existing scales (the Rapid Risk Assessment for Sex Offence Recidivism; Thornton's Structured Anchored Clinical Judgement) and contains the following items: prior sex offences, unrelated victims, stranger victims, male victims, age, never married, non-contact sex offences, four or more prior sentencing dates, current non-sexual violence, and prior non-sexual violence.

The scale was tested in four diverse samples of sex offenders drawn from secure psychiatric institutions in Quebec (n = 344) and Ontario (n = 142), an Ontario provincial prison (n = 191) and from Her Majesty's Prison Service (England and Wales)(n = 531). Recidivism information was based on national police records, except in one sample that used both law enforcement and hospital records. For most offenders, the follow-up period was greater than 15 years.

Answer

Averaged across the samples, the recidivism rates after 15 years were 26% for sex offences and 37% for any violence (including sexual). Static-99 showed moderate accuracy in predicting both sexual and violent recidivism. After 15 years follow-up, the sex offence recidivism rate of the low risk group was approximately 10%, whereas the recidivism rate of the high risk group was greater than 50%. The recidivism rate for any violence was 17% after 15 years for the low risk offenders, compared to 59% for the high risk group.

The degree of predictive accuracy was similar across the four samples, showing no more variability than would be expected by chance. As well, the scale was equally accurate for those who had offended against children (child molesters) or adults (rapists).

Policy implications

  1. Structured risk scales have sufficient accuracy in predicting sex offence recidivism to be useful in applied contexts. The predictive accuracy is far from perfect, but it is clearly superior to other methods commonly used to assess risk with sex offenders (e.g., unstructured clinical judgement).
  2. The ability to distinguish between low risk and high risk sex offenders is sufficient to support special policies directed at high risk offenders (e.g., Long Term Offender designations, intensive treatment programs).
  3. Most sex offenders are never reconvicted for a sex offence, and less intensive interventions (e.g., probation) can be provided to low risk offenders without compromising public safety.

Source

For further information

James Bonta, Ph.D.
Solicitor General Canada
340 Laurier Avenue West
Ottawa, Ontario
K1A 0P8
Tel (613) 991-2831
Fax (613) 990-8295
e-mail jim.bonta@ps-sp.gc.ca

Date modified: