ABSTRACT
In the previous three chapters, we have presented a brief history of actuarial assessment of risk for recidivism in sex offenders. We described how actuarial risk assessment instruments (ARAIs) were developed and illustrated our description with the development of our own brief (three-item) actuarial scale for child molesters. We described how ARAIs are evaluated for accuracy and we surveyed the literature focusing on independent cross-validations of several ARAIs. We examined the problems experienced in calibrating ARAIs, specifically the fact that estimates of score-wise recidivism rates are characterized by large error rates. In this chapter, we examine some other problems with ARAIs, and make the claim that many of these problems arise because these instruments were developed in the absence of any theory of recidivism risk. We underscore the importance of theory in science, and we speculate on what solutions to problems might be found with the guidance of a theory of risk.
