Problems in Modeling and Measuring Recidivism

Abstract

Despite the fundamental character of recidivism as a measure of correctional effectiveness, and its wide use in all aspects of the theory and operations of the crime control process, there is sur prising misconception regarding its use and measurement. This paper uses probabilistic models to examine two types of potential misconception. The first arises from alternative definitions of recidivism based on the stage in the criminal justice process at which recidivism is measured. The second arises from alternative methods of sampling criminal careers. The analysis helps to clarify possible interpretations of recidivism data, including those reported by the FBI under "Careers in Crime."