The reliability and stability of self-reports of licit and illicit drug use are estimated using longitudinal data, and found to be rather high. An apparent inconsistency in reporting (less use over a 12-month period than would be expected based on use reported over a 30-day period) is examined (1) for association with personal characteristics and (2) for individual consistency over time. Neither individual consistency nor significant associations are found, although there appears to be a tendency for discrepant reporting of marijuana and alcohol use to be correlated within time. The data used are from the Monitoring the Future project, an ongoing nationwide study of high school seniors, with follow-up surveys after graduation.