This research investigated the use of telephone versus face-to-face interviewing to gather data on the use of tobacco, alcohol, and illicit drugs. Telephone and personal drug use surveys of the 18-34 year-old household population were conducted in the state of New Jersey in 1986-1987. Survey modes were compared in terms of unit and item nonresponse rates, sample coverage, and levels of self-reported drug use. Results showed that the telephone survey achieved response rates lower than the personal survey, but comparable to telephone surveys of less threatening topics. Item nonresponse to sensitive drug questions was lower by phone than with the self-administered answer sheets in the personal mode. The exclusion of households without telephones in the telephone survey is a potential source of bias, and may lead to underestimation of alcohol and drug use for minority populations. After controlling for telephone status, the telephone survey furnished significantly lower drug use estimates on several indicators than the personal survey, with the largest mode differences found for Blacks.