DAWN seeks to monitor the number of drug-related emergency room (ER) episodes by retrospectively examining hospital ER records. The sample consists of a subset of the nonfederal, general-care, short-stay hospitals that operate a 24-hour emergency department. The current probability sample is an improvement over the previous convenience sample. Although DAWN serves some purposes well, it is not appropriate for other purposes. In addition, more than a passing knowledge of DAWN is needed to distinguish between appropriate and inappropriate applications. Although DAWN seeks to provide information about a particular definition of drug-related ER episodes, its estimates may differ from the true number and in ways that vary over time and place. In addition, a count of drug-related ER episodes can provide only limited information about the size and composition of the drug-using population or the overall health consequences incurred by that population. Therefore, policy analysts need to be more circumspect about the conclusions reached from DAWN data. Notes and 35 references