Recent studies investigating the labor-market effects of illicit drug use have consistently found a positive relation between drug use and earnings. These analyses have, however, ignored the potential relationship between drug use and human-capital formation. This paper examines the effect of drug use during adolescence on formal educational attainment using a sample drawn from the National Longitudinal Survey Youth Cohort. The probability of drug use is estimated across racial groups according to three categories; use of any illicit drug, use including hard drugs, and use of only marijuana. Fitted values for the probability of drug use are calculated and entered into a regression framework to estimate the number of school years completed. The empirical results indicate that all three categories of drug use are associated with significant negative impacts on educational attainment after controlling for individual differences in personal endowments and socioeconomic characteristics. On average, adolescent drug use is found to reduce eventual educational attainment by about 1 year, ceteris paribus . These findings suggest that previous studies that focus only on the direct effects of drug use on earnings may reflect a statistical bias that leads to an overstatement of the positive effects of drug use on earnings.