The concept of a "criminal career" is a useful means of characterizing the longitudinal sequences of individual criminal activity. It can be done for a particular individual or for a population of individuals who engage in crime. Much of the research on criminal careers has involved a theoretical measurement of career patterns and parameters. One of the most important areas of measurement is offense frequency; in criminal-career modeling, this variable has been symbolized by the Greek letter lambda. In the simplest model of a criminal career, the individual starts at an initiation age and continues until the termination age, committing crime at a rate (lambda) over the duration. In addition to frequency of offending (lambda), variables that should be considered in modeling criminal careers are the probability of arrest following the commission of a crime, individual arrest frequency, the probability of incarceration given arrest, the probability of incarceration given a crime, and the time served in incarceration when a sentence is imposed. Regarding policy implications of criminal-career research, some have proposed the selective incapacitation of persons in the early stage of crime for those who fit the parameters for long-term criminal careers. The unreliability of prediction instruments and the sentencing disparity involved, however, have doomed this policy recommendation. Other research has suggested that for those who engage in a high frequency of crime commission, there is a natural selective incapacitation because they are more likely to be arrested and incarcerated than those who commit relatively fewer crimes. Some criminal-career research needs are better assessment of the shape of the criminal career patterns, more attention to factors that affect crime frequency or career duration, identification of the effects of incarceration on criminal career patterns following release, and examination of switching between crime types in the course of a criminal career. 6 figures, 2 notes, and 18 references