A total of 202 male dropouts, mainly heroin abusers averaging 3 to 6 years out of treatment, were sampled from the 1970-1971 Phoenix House (New York) admissions by race and time in program (greater than 1 month and less than 2 years). Criminal justice arrest records were compared between three pre- and post-treatment years. Percent and rate of arrest were shown to decline significantly in followup. The magnitude of reductions was uniform across race and legal status, varied by age, but increased systematically by the time in the program. Results replicated and extended earlier findings, confirming long-term positive change in criminal behavior associated with length of stay in the therapeutic community. Preliminary analyses further show that reductions in arrests are (1) correlated with reduced drug use and improved employment, (2) occur similarly for drug- and nondrug-related charges, (3) correlate positively with self-reported criminal activity, and (4) are client-specific (reductions are maintained among those who curtail criminal activity). In a long-term followup, however, the influence of other treatment and nontreatment factors must be assessed in terms of their influence on the reduction of criminal activity. External influences (the availability of heroin on the street, law enforcement pressures, entry into other drug programs, the economy, employment, etc.) as well as client-related factors (motivation, maturation) cloud precise identification of the effects of treatment at Phoenix House. The close replication of other findings, however, supports the effectiveness of the therapeutic approach in modifying antisocial behavior among drug abusers. A total of 18 references and tabular and graphic data are provided. (Author abstract modified)