This paper examines the validity of self-report of drug use in several important high-risk samples by comparing survey answers with urinalysis results. The analyses include data obtained from face-to-face interviews with 1536 patients in sexually transmitted disease (STD) clinics, 1564 patients in hospital emergency rooms (ERs), and 2034 recent arrestees in jails. Using urinalysis as the criterion measure, considerable underreporting was observed, particularly among STD and ER subjects using cocaine and opiates. A logistic regression analysis on factors related to the underreporting was conducted for both cocaine and opiates. The results showed that a lower degree of underreporting was associated with subjects at jail sites (compared to STD subjects) and those who self-reported lifetime drug dependence.