Graduation rates in drug courts average 50% to 70%, but it is unclear what proportion of graduates responded to the drug court services and what proportion might not have had serious drug problems upon entry. This study cluster-analyzed urine drug screen results during the first 14 weeks of treatment on 284 participants from three misdemeanor drug courts. A four-cluster solution (R(2) > .75) produced distinct subgroups characterized by (1) consistently drug-negative urine specimens (34% of the sample), (2) consistently drug-positive specimens (21%), (3) consistently missed urine specimens (26%), and (4) urine specimens that began as drug-positive but became progressively drug-negative over time (19%). These data suggest that approximately one-third of the participants might not have had serious drug problems upon entry. Approximately one-fifth appeared to respond to drug court services, and nearly one-half continued to exhibit problems after 14 weeks. Implications for adaptive programming in drug courts are discussed.