Perhaps the first of its kind in the early 1960s, this study explored the methods and challenges of identifying medication errors. Targeting registered nurses as a source for data collection, the investigators studied direct observation, self-report, and existing incident reports to compare strategies for effective error detection. The authors review the definition and types of medication errors before presenting findings from their contrasting data collection methods. Conclusions suggest that direct and disguised observation may be the best method for estimating medication errors, and that previous efforts to assess the prevalence of the problem are likely understated.