Results show that many of the trends in offense specialization across arrests are likely influenced by changes in specialization that occur with age. The findings show that controls for age did not have a significant influence on trends in offense characteristics across arrests for drug offenses or miscellaneous offenses; however, controls for age did influence trends in specialization across arrests for violent offenses and property offenses. For both of these offense types, trends in offense specialization across arrests prior to controls for age were positive; and trends in offense specialization across arrests after controls for age were negative. This article discusses the theoretical implications of these findings, as well as methodological considerations and extensions of the current work. The analysis used data from the study entitled, "Predicting Parole Performance in the Era of Crack Cocaine" (Haapanen and Skonovd, 1999). Participants were in custody under the supervision of the California Youth Authority during the 1980s. Data were drawn from a random sample of 2,086 offenders released from custody in fiscal year 1981-1982. Data were obtained from four sources: Youth Authority electronically stored data files, Youth Authority hard-copy master files, California Department of Justice criminal history files, and California Vital Statistics. The sample was composed of serious offenders with many development and contemporary correlates of crime, including high rates of abuse or neglect, drug and alcohol use, and gang association. At the end of the data-collection period, offenders ranged in age from 24 to 35 years. Data included both juvenile and adult offense histories. Arrest offense types were coded into violent, property, drug, and miscellaneous offenses. 6 tables and 42 references