Commentary

Comments on five analyses (in this issue) of three surveys of US youth substance use prevalence rates that found, overall that differences across the surveys is mainly attributable to administration mode. Research design (eg, replacement procedures) differences & the impacts of nonresponse, anonymity, & parental consent are briefly considered, along with the potential for the disparate handling of inconsistent & missing data to generate disparate prevalence rates. Also, limitations in survey documentation are sited as a factor in impeding interpretation of findings. Evaluation of similar ongoing surveys is deemed a means of aiding the assessment of differential results. 4 References. M. C. Leary