Abstract
Despite efforts to incorporate protective factors or 'strengths' in applied risk assessments for criminal reoffending, there has been limited progress towards a consensus regarding what is meant by such terms, what effects predictors can exert, or how to describe such effects. This proof of concept study was undertaken to address those issues. A structured professional judgment tool was used to create lower and higher historical/static risk groups with a sample of 273 justice-involved male youth with sexual offenses followed over a fixed 3-year period. Using risk and protective poles to create pairs of dichotomous variables from trichotomously rated risk and protective items, risk-based exacerbation and risk-based protective effects were found. These varied in terms of whether the effect on the outcome of a new violent (including sexual) offense was larger, smaller, or absent for youth at higher or lower historical/static risk. Some of these potentially dynamic dichotomous variables were shown to have a protective (or risk) effect after controlling for both historical/static risk and that same item's risk (or protective) effect. Some moderated the association between historical/static risk and recidivism, strengthening or reducing it. Terms for these effects and implications of incorporating strengths in research and applied practice were considered.