Prison and probation risk assessment tool is not effective in judging re-offending

A psychological risk assessment tool used by prisons and the probation service (NOMS) is not effective when evaluating the future risk of convicted male sex offenders.

In a new paper published online in Criminal Justice and Behaviour, Dr Ruth Tully and Professor Kevin Browne from the Centre for Forensic and Family Psychology at The University of Nottingham, together with Professor Leam Craig from the University of Birmingham, examine the effectiveness the SARN-TNA (Structured Assessment of Risk and Need – Treatment Needs Analysis) in predicting the reconviction of sex offenders.

The SARN-TNA has been used routinely by NOMS, and has been heavily relied upon to assess sex offenders' need for treatment and whether they should be released early from prison. When a sex offender is released from prison, probation services often use this tool to decide on the type and intensity of restrictions, treatment, supervision and surveillance that the sex offender receives.

Ineffective in decision-making

External bodies such as the parole board heavily weight the SARN TNA assessment in its decision-making, for instance when deciding to release prisoners serving a life sentence. These assessments are so heavily relied upon that parole board hearings are often deferred at costs of thousands of pounds to the taxpayer if a SARN-TNA has not been conducted - but until now the SARN-TNA has not been properly or independently tested. The SARN-TNA has now been shown to be ineffective.

These were the alarming findings of a four-year study of reconviction within a population of 496 adult male sex offenders, who were assessed using the SARN-TNA. The SARN-TNA assigns a risk level of low, medium or high to each prisoner. The study found that the SARN-TNA risk levels demonstrated no predictive accuracy at a two and four year follow-up.

When the tool was examined to see if it significantly predicted reconviction in any way, only one of its four domains (sexual interests) was found to be predictive of sexual reconviction. Additionally, reconviction rates were not significantly different between risk groups; meaning that 'high' risk men did not differ in rate of reconviction compared to 'low' risk men.

First study of its kind

Dr Tully says: "This field based study is the only study of its kind in the UK. It found that the SARN-TNA is not effective in predicting risk of sexual reconviction. This is particularly concerning when offenders are released on the basis of opinions of risk reduction using an ineffective framework. Probation Services aren't able to effectively manage sex offenders if they haven't been appropriately risk assessed. As a worst case scenario, if this tool is relied upon, a 'low risk' sex offender whose risk was really 'medium to high' could be released from prison early and reoffend.

"Forensic risk assessment by psychologists is complex and difficult, and defensible decision making is crucial. A 'one-size-fits-all' approach of applying the SARN-TNA to assess risk after treatment, regardless of the specifics of the case or effectiveness of the tool is not acceptable. Sex offenders should not be released on the basis of a tool that is statistically no better at predicting reconviction than chance.

"Clinicians and policy makers should consider very carefully the future of SARN-TNA as a risk and treatment need assessment given the findings of this study. Agencies should take a critical approach to the use of this tool, and also carefully consider the use of alternative sex offender risk assessment tools. National Offender Management Services (NOMS) should routinely allow independent bodies such as Universities ease of access to their data to conduct research. This would aid advancements in knowledge in the field of sex offender risk assessment."