Five of the 14 ISP demonstration sites sponsored by the Bureau of Justice Assistance designed and implemented programs specifically for drug offenders, focusing mostly on surveillance rather than treatment and using group counseling as the main form of treatment. At each site, nearly 600 adult drug offenders were randomly assigned to ISP or routine probation/parole and tracked for a year from the date of assignment. The RAND evaluation of these programs revealed that ISP apparently did not affect drug use, did not reduce recidivism, and cost more than routine supervision. However, the ISP's were successful as intermediate sanctions. The results raise two issues: (1) whether making ISP tougher would produce more crime control and (2) whether an increased emphasis on rehabilitation would make ISP more effective by changing offender behavior. Because the demonstration programs that were not focused on drug offenders had similar results and because most serious offenders are drug-involved, research is needed that addresses these and related issues. Figures and 7 references