Because mailed surveys minimize personal contact, they are useful for collecting sensitive data on substance use, as long as the problems of achieving adequate response rates can be conquered. To address these issues, we report on an anonymous mailed survey of substance use with a 78% response rate, including data collection and survey methods. Analysis of sociodemographic effects on responding found certain groups required additional contacts. Substance use estimates were not affected by non-response bias, suggesting that anonymous mailed surveys can be a feasible means of collecting data on substance use.