Kirk Heilbrun put forward the prediction and management models in 1997. The prediction model refers to an assumption of potential future behaviours among offenders, whilst the management model allows for hopefully solutions to those undesirable behaviours (Heilbrun, 1997). This essay will focus on methods of assessing sexual offenders, by identifying and evaluating the most commonly used assessments, and identifying where these methods are in relation to the Prediction and Management models.
Focusing on sexual offenders is important because sex crimes appear to be on the rise, climbing by 10 percent (Ministry of Justice, 2015). Alongside this, sex offenders who are left untreated when they leave prison have been shown to have very high rates of recidivism (Hanson and Bussiere, 1998). By potentially identifying both those at highest risk of reoffending or are at higher risk of offending to begin with, secondary effects could go as far as lowering the amount tax payers have to pay for incarcerating such high levels of offenders. As of current, tax payers pay inflated amounts towards the prison system (Henrichson and Delaney, 2012).
The Sex Offender Risk Appraisal Guide (SORAG) is a scale consisting of 14 items, designed to predict the likelihood of re-offending of criminals who have been previously convicted of sex crimes at least once before (Quincy, Harris, Rice & Cormier, 1998). SORAG focuses on static factors, factors which cannot be changed. This includes things such as criminal history, relationship with parents, and young school life.
On multiple occasions, the SORAG has been found to be effective at predicting recidivism rates (Ducro and Pham, 2006; Barbaree, Seto, Langton & Peacock, 2001). However, it’s been argued that SORAG is more effective at predicting those are who are less like to relapse than those who are (Eher, Reinhard, Schilling & Pfafflin, 2008). Furthermore, the SORAG bases its predictions on a singular score created by weight placed on each of the 14 items, which is then added at the end and used to determine the likelihood of re-offending. The problem with this may be that using a singular score may be of minimal value when trying to establish the chances of reoffending because it provides so little information (Hart, 1998).
Another problem that somewhat stems from predicting from a singular score is that it makes it difficult to predict specific sexual offender groups, such as child sex offenders, indecent exposure, rapists. In fact, Harris and Rice (2003) found that the SORAG was unable to predict relapse for rapists. This inability to specifically evaluate subgroups of offenders is a problem with many actuarial scales, including for instance the Static-99 scale (Hanson and Thornton, 2000). Much like the SORAG, the Static-99 was found to be good for general recidivism predictions, but poorer for specific judgements. This could cause problems if a specific offender whose higher at risk of re-offending is released because the general score comes out as low risk
A comparison between the SORAG and Static-99 found that both measures looked at the same information, thus meaning that the measurements may provide useless or redundant information (Nunes, Firestone, Bradford, Greenberg & Broom, 2002). This implies that the SORAG could not be improved even by combining it with the Static-99, which overall brings its utility into question.
The utility of the SORAG is minimal- for a base score, it works well but it provides little help beyond that. Not only does it lack specific predictability, but there is no sort of follow up or recommended solution to that risk. It does not fulfil the management model at any level, which means that it’s usefulness in the real world does not extend far. In fact, it is arguable that there is no use for the SORAG- if the Static-99 covers all the same bases, and neither provides any sort of management solution, then the SORAG serves no unique purpose in the field of risk prediction and management (Nunes, Firestone, Bradford, Greenberg & Broom, 2002).
Further reducing it’s applicability in the real world is the fact that the SORAG was specifically designed to measure risk of re-offending among men (Barbaree, Seto, Langton & Peacock, 2001). Therefore, whether or not it is a suitable measure for female sex offenders has not been explored, which means the SORAG’s utility is even further restricted.
However, there is grounds to question whether or not the SORAG is measuring what it is supposed to. Hanson and Bussiere (1998) found that 5 out of 14 of the items on the SORAG have little to no correlation with violent sexual crime. This brings the validity of the SORAG into question. There may be other, irrelevant factors in the list which further reduces the validity, and makes the SORAG more useless in terms of utility.
Finally, the SORAG also uses the Psychopathy Checklist-Revised (PCR) scoring in it's overall assessment. The PCR is mostly objective, but there's areas of it which require clinical judgement. These areas are so subjective that specialist training is required in order to use the PCR correctly (Wakefield and Undeerwager,1998). If it's misused or misjudged, the evaluator risks breaking one of the ethical codes. Since the PCR is a part of the SORAG, misuse of the SORAG may result in the same consequence . Furthermore, because of how subjective those areas are, difficulty in accurately predicting recidivism is likely to be high, unless each prediction was found to have high interrater reliability, which due to time and confidentiality constrictions, may prove impractical.
In conclusion, the SORAG is continually accurate and effective at predicting general reoffending rates, which fits in the Prediction model. However, prediction is limited to general, and SORAG is unable to predict subgroups of offenders specifically. This, combined with it’s limited result of one value, applicability to only men, and it’s s questionable validity means that the SORAG is not ideal for use in the real world. This is further outlined by it’s lack of management suggestions, which limits utility even more, possibly even making it redundant.