Rediscovering Quetelet, again : the “aging” offender and the prediction of reoffending in a sample of adult sex offenders

This study explored the role of age at release on the risk of reoffending using a sample of sex offenders. It examined whether the risk of reoffending, assessed using actuarial tools, should be adjusted according to the offender’s age at the time of release. The sample comprised 553 offenders, all of whom were consecutive admissions to a Canadian federal penitentiary. Scores on the Static-99 as well as age at release were included in successive nested prediction models using Cox-regression. Receiver operating characteristic (ROC) curves and Allison’s R 2 were computed to assess the predictive accuracy of the models and the strength of the association between the covariate measures of general and violent/sexual reoffending. Results showed that overall predictive accuracy observed across models was fair at best. Generally, age of onset and age at release improved the prediction accuracy over and above the scores on the Static-99. In fact, by itself, age at release showed a predictive accuracy comparable to that of the actuarial tool. The results suggest that risk assessors should adjust the risk of reoffending based on the offender’s age at release. The implications of this study are discussed in light of the age–crime curve literature and the risk management of sex offenders in the community