The drug court involved in the study was located in a mid-sized northeastern city. It conformed to the model of drug courts that has been promulgated and supported by the Drug Courts Program Office of the U.S. Department of Justice. Most of the data used in this study were obtained from a database created and maintained for the monitoring and evaluation of the "City Drug Court." This involved records of all persons enrolled in the court since its inception in 1997 through April 30, 2001. The universe for the analysis consisted of the cohort of drug court participants enrolled from March 1997 through March 31, 2001 (n=397). Information obtained included participants' demographic characteristics, current charges, critical participation-related dates, drug of choice, and type of representation. A second database provided several variables for the analysis, including criminal history, current employment status, and educational attainment. The dependent variable measured outcomes (continue, prematurely terminate, and graduate) for each participant for each month that he/she was enrolled in the court. The analysis identified factors associated with both successful completion of the drug court program and premature termination. The study found that having an African-American case manager, being older, having little criminal history, and not using crack cocaine were strongly predictive of successfully completing the program. A low probability of successful completion of the program was found for young non-White participants who used crack cocaine as their primary drug and who were assigned to a White case manager. 3 tables, 1 figure, and 31 references