Criminal Careers: Discrete or Continuous CANCELLED

Numerous empirical studies of criminal careers have made use of finite mixture modeling to analyze sequences of events such as crimes or arrests. This approach represents a set of individual event sequences with a finite number of discrete trajectories. Individuals are tyopically matched to the trajectory that is most likeley, given the disctinctive pattern of that individual's sequence. We use hierarchical linear modeling and individual time series techniques to test the assumption that arrest trajectories are meaningfully and usefully classified into a small number of mutually exclusive discrete classes, using data for 332 males released from the California Youth Authority in 1981 and 1986, and followed for several decades after release. In this data set we find little evidence of sharply discrete arrest trajectories.