ACA Research Council Sponsors Author-Meets-Critic Session

Article excerpt

The corrections field is at a crossroads as it struggles with fiscal pressures and the need for informed resource allocation to ensure the use of cost-efficient and cost-effective strategies. During the past three decades, the field has moved from using relatively unstructured risk assessment to implementing more systematic and empirically validated approaches that inform case management both in prisons and in the community. Ideally, correctional systems should use a model that borrows from each of these approaches and allows for the sharing and integration of key case management information in decision-making.

In February 2009, the National Council on Crime and Delinquency (NCCD) issued a special report titled "A Question of Evidence: A Critique of Risk Assessment Models Used in the Justice System." (1) Its author, Christopher Baird, is the executive vice president of NCCD. The report raises a number of important issues regarding contemporary risk assessment approaches and, in the process, sparks a flurry of debate. As it has done in the past with other noteworthy and controversial reports, ACA's Research Council sponsored an author-meets-critic session to provide a forum for discussion. The session took place at the 2010 ACA Winter Conference in Tampa. I was the "critic" to Baird and, at the request of the National Institute of Corrections, which provided financial support for the session, I wrote a report that expounds on my remarks in Tampa and draws from my comments below.

A major premise of the NCCD report is that the predictive accuracy of risk scales can be enhanced by addressing the key aspects of scale development more rigorously. Scale items (both static and dynamic) vary in predictive validity, but it is well-known that knowledge of only a small number of criminal history and demographic variables can yield modest to good predictive accuracy. The NCCD report argues that risk assessment accuracy diminishes when too many risk and need items are calculated in the same instrument and that increasing the number of items in a risk scale is problematic because all the items are not equally predictive. However, many agencies use risk assessment both to assign risk levels (e.g., for classification) and to develop program plans; so, a combined scale has appeal. Unfortunately, a recent meta-analytic review of risk scales shows no marked difference among scales used commonly in correctional and forensic assessment. Thus, combining the scales does not strengthen them or make them any more effective. (2)

The NCCD report also challenges the definition of "criminogenic," asserting that, at a case level, many factors (e.g., self-esteem) could cause criminality. But this seems to contradict the consistent findings from meta-analytic research that there are essentially seven or eight major criminogenic domains. While it is possible to describe a case for which self-esteem is a credible factor, making it part of formal programming will surely erode the gains made in terms of the development of the theory of offending behavior and the cost-effectiveness of correctional programs. This is not to say that the field should be closed to new understanding about antecedents to criminal behavior, because new domains, such as trauma and mental health symptoms for female offenders, are emerging. (3)

The report also notes the need to be more judicious in applying the term "protective factor," because not doing so may raise false expectations. Crime desistance, which has been described in the research literature for more than two decades, is gaining popularity as researchers increase their efforts to describe and measure constructs underlying offenders' transition to successful reentry. Because of this, increased specificity is required, particularly in the area of protective factors, which inform the crime desistance process. As the report correctly notes, protective factors are not simply the absence of risk factors, and it is the field's obsession with risk assessment that has obscured its understanding of offender success. …