A random sample of 1,200 employees of a steel plant in the western United States was randomly assigned to four different self-report methods of assessing illicit drug use: individual interview in the workplace, group-administered questionnaire in the workplace, telephone interview, and individual interview off the worksite. Urine specimens were collected and analyzed on all 928 subjects participating in the study, and hair analysis was conducted on 307 of the subjects. Although self-reports produced higher prevalence rates than the chemical tests, analyses combining the results of the three assessment methods showed that the actual prevalence rate was approximately 50 percent higher than the estimate produced by selfreports alone. The group-administered questionnaire method produced prevalence rates that were roughly half those of the other self-report methods. The findings cast doubt on the validity of selfreports as means of estimating drug use prevalence and suggest the need for multiple assessment methods.