Background-Recent media reports have portrayed an alarming increase in apparent anabolicandrogenic steroid (AAS) use among American teenage girls; Congress even held hearings on the subject in June 2005. We questioned whether AAS use among teenage girls was as widespread as claimed.Methods-We reviewed four large national surveys and many smaller surveys examining the prevalence of AAS use among teenage girls. Virtually all of these surveys used anonymous questionnaires. We asked particularly whether the language of survey questions might generate falsepositive responses among girls who misinterpreted the term "steroid." We also reviewed data from other countries, together with results from the only recent study (to our knowledge) in which investigators personally interviewed female AAS users.Results-The surveys produced remarkably disparate findings, with the lifetime prevalence of AAS use estimated as high as 7.3% among ninth-grade girls in one study, but only 0.1% among teenage girls in several others. Upon examining the surveys reporting an elevated prevalence, it appeared that most used questions that failed to distinguish between anabolic steroids, corticosteroids, and over-the-counter supplements that respondents might confuse with "steroids." Other features in the phrasing of certain questions also seemed likely to further bias results in favor of false-positive responses.Conclusions-Many anonymous surveys, using imprecise questions, appear to have greatly overestimated the lifetime prevalence of AAS use among teenage girls; the true lifetime prevalence may well be as low as 0.1%. Future studies can test this impression by using a carefully phrased question regarding AAS use.