Research and analysis

Actuarial prediction of sexual reoffending: responding to changing offending patterns

This paper analyses HMPPS actuarial sexual reoffending prediction tool, looking at the causes of its reduced performance and identifying solutions.

Applies to England and Wales

Documents

Annexes

Policy addendum

Details

These papers constitute an investigation into the actuarial instruments, designed by the MoJ and used by HMPPS, to estimate individual offenders’ risks of proven reoffending.

The first paper documents work done to analyse the risk of reoffending across several different offence types. All Actuarial Risk Assessment Instruments currently in use by MoJ are revalidated, assessing their ability to both accurately predict rates of reoffending for groups of offenders and distinguish between individuals with high and low risks of reoffending.

The second paper looks specifically at the OASys Sexual Predictor (OSP), the actuarial instrument used to predict sexual reoffending. As the first paper identifies a reduction in OSP’s predictive validity – although this is still at acceptable levels – the second paper seeks to understand what may be causing this, looking at recent changes in offending patterns, and identifying potential changes to OSP’s algorithms that would remedy this issue.

Finally, the policy addendum notes the policy changes which are to be undertaken by HMPPS in response to the papers published

Revalidation: Risk of recidivism tools, an evaluation of the actuarial instruments developed to assess recidivism risk in England and Wales

Updates to this page

Published 1 February 2024

Sign up for emails or print this page