California's voter-initiated Proposition 36 offers non-violent drug offenders community-based treatment as an alternative to incarceration or probation without treatment. This article reports short-term treatment outcomes subsequent to this major shift in drug policy. Data are from 1104 individuals randomly selected from all Proposition 36 participants assessed for treatment in five California counties during 2004. The overall study sample was 30% female, 51% white, 18% Black, 24% Hispanic, and 7% other racial/ethnic groups. The mean [plus or minus] SD age was 37 [plus or minus] 10 years. Counties varied considerably in participant characteristics, treatment service intensity, treatment duration, urine testing, and employment and recidivism outcomes, but not in drug use at 3-month follow-up. Controlling for county, logistic regression analysis showed that drug abstinence was predicted by gender (female), employment at baseline (full or part-time), residential (vs. outpatient) stay, low psychiatric severity, frequent urine testing by treatment facility, and more days in treatment. Recidivism was predicted only by shorter treatment duration. Employment predictors included age (younger), gender (male), baseline employment, and lower psychiatric severity. The study findings support drug testing to monitor abstinence and highlight the need to address employment and psychiatric problems among Proposition 36 participants. (Contains 3 tables.)