Testing for the use of marijuana necessitates one or more levels of assessment, depending on the clinical application of the test results. The EMIT-dau screening test with a sensitivity limit of 20 ng/ml is highly satisfactory for screening for the absence of marijuana in the urine. This test has a virtual 100% true-negative rate as long as an unadulterated urine specimen is analyzed. Positive results by the EMIT-dau procedure are presumptive because the test seems to produce false-positive results when applied to a random population of suspected drug users. In a population in a specific clinical environment such as in drug-treatment programs in which 20 of every 100 specimens will yield positive THC-COOH results by EMIT-dau screening, 3 (15%) of the 20 positive results will likely be false-positive. Our experience with the EMIT-dau suggests that of 100 test results, 3 will be false-positive, an overall 3% false-positive rate. If a positive test result will put the patient in considerable jeopardy and the screening result is the only evidence of drug use, confirmatory testing is imperative. Of the confirmatory tests, GC/MS seems to have the specificity necessary to provide a high level of confidence in the results. A combination of the EMIT procedure with a sensitivity level of 20 ng/ml and GC/MS confirmation yields virtually 100% accuracy in detection of marijuana abuse.