To investigate the possibility that self-reported smoking is not a valid measure for assessing trends in smoking prevalence, we compared total self-reported cigarette consumption with the adjusted consumption data from cigarette excise taxes as reported by the US Department of Agriculture (USDA) for the period 1974 through 1985. Self-reported consumption was calculated by using data from the National Health Interview Surveys (NHIS) for adults and from the National Household Surveys on Drug Abuse for adolescents. For this period, the average ratio of self-reported cigarette consumption to the USDA estimate of consumption was 0.72 (range = 0.69 to 0.78). There was no statistical difference in this consumption ratio from year to year, indicating no apparent increase in the underreporting of cigarette smoking in these surveys. We conclude that cross-sectional surveys of self-reported smoking status remain a reliable surveillance tool for monitoring changes in population smoking behavior.