Excerpt from the This volume contains contributions representing some of the major directions of recent modeling efforts, beginning with a focus on criminal (or delinquent) careers and their restraint. Chapter 1 by Greene and Stollmack contains a basic attempt to estimate the size of the offender population. In Chapter 2 Figlio also employs a probability distribution to describe delinquent career patterns, in particular a Markov model for the transition from one offense state to the next. In Chapter 3 Blumstein and Nagin use assumptions concerning criminal career patterns to explore the efficacy of various crime control strategies. In Chapter 4 Harris, Kaylan, and Malta review and contrast several of the probability distributions that have been advanced for modeling the recidivism process. In Chapter 5 Barton and Turnbull avoid the distribution issue in assuming a nonparametric approach to failure rates. As a next step, they develop regression models for recidivism, adopting Cox’s regression approach to censored failure times with extensions for grouped data. Most significantly, Barton and Turnbull demonstrate how offender-specific covariates in a failure rate regression model can incorporate the important notion of base expectancy in order to enhance comparative evaluation efforts.