Inspired by: Mann, R.E., Hanson, R.K., & Thornton, D. (2010). Assessing risk for sexual recidivism: Some proposals on the nature of psychologically meaningful risk factors. Sexual Abuse: A Journal of Research and Treatment. Published on April 2, 2010 as doi:10.1177/1079063210366039.
I've been somewhat consumed of late regarding this issue of how best to compare certain offenders into the various Static-99R normative samples. Of course, this process has huge implications, both for the offender himself and for community safety. The process of choosing between “routine”, “preselected for treatment participation”, “preselected for risk/need”, and “non-routine” has generated healthy debate (see Campbell & DeClue, 2010--link 1 below; Wilson & Looman, 2010--link 2 below), but no clear answers or protocols as to best practices.
I have argued elsewhere (link 2) that most evaluators will accept that there are risk factors that are important to consider--but not necessarily precisely tapped by Static-99R--in comprehensively determining risk to reoffend. These factors might include sexual deviance, psychopathy, and other factors as delineated in the literature. However, there are as yet few systematic protocols for combining these factors with static actuarial scores. During a recent preconference workshop at the annual ATSA conference in Phoenix, David Thornton stated:
For any given Static-99R score common degrees of variation in Need produce dramatic differences in recidivism risk. This remains true within any specified degree of pre-selection. You can’t accurately evaluate recidivism risk without systematically evaluating Need.
I couldn't agree more and, as professionals in the field, I encourage you all to read the paper noted above by familiar ATSA members Ruth Mann, Karl Hanson, and David Thornton. In this highly readable paper, the authors present an excellent summary of the risk assessment process, reviewing various categories of risk factors and the many means by which to measure them. Overall, I found their perspective refreshingly honest. Two of the authors (Karl and David--and I'm certainly not suggesting that Ruth is not also a big part of this field) clearly have a lot at stake, given their respective contributions to the risk assessment literature. Yet the paper freely admits that many methods are less than perfect and that we still have much to learn.
One of my favorite lines in the paper came early (in the abstract):
Although it is possible to conduct risk assessments based purely on empirical correlates, the most useful evaluations also explain the source of the risk. (emphasis in the original)
I have read many a risk assessment in my time and, unfortunately, all too often there is a certain mechanical quality to them. I like the idea of making sure that the receivers of our risk assessment findings can be better informed as to what we're actually talking about, especially if this means including elements of elegance and eloquence along with the facts and data. As such, the term "psychologically meaningful" resonates with me.
Mann et al. make distinctions between risk factors that are empirically supported, promising, unsupported--but with interesting exceptions, worth exploring, and those with little or no relationship to sexual recidivism. They provide suggested means by which to ascribe factors to these various categories, as well as examples of each (with a comprehensive review of the literature underscoring each). This latter aspect makes the paper quite useful to persons who don't necessarily have the time to review each and every individual paper on factors potentially predictive of risk to reoffend. Ultimately, they turn to the concept of "causal" factors--those that might actually be so pertinent as to provide reasons for offending. Mann et al. state that "assessment and treatment for sexual offenders should focus on empirically established causal risk factors," but stop short of telling us what those factors are. That's where our research needs to focus going forward.