Eight hundred fifty-four high school students' responses were employed to compare drug-use estimates derived from either (a) traditional direct questioning regarding use of six drugs, or (b) a more indirect and more anonymous method of inquiry, the randomized response procedure. Results indicated that (1) the randomized response procedure produced significantly fewer response refusals, and (2) significantly higher drug-use estimates supported the hypothesized greater sensitivity and validity of the randomized response procedure. The results further suggested that previous estimates derived from standard forms of questioning many have underestimated incidence of drug-use.