To the Editor: Three flaws in the study by Kralewski et al. (July 28 issue)1 render its hypotheses questionable, reduce the internal and external validity, and do not permit the data to support the conclusions drawn. First of all, the study is based on the notion that pharmacists' drug-product–selection behaviors can be explained by using two categories of drugs: “generic drug products” and “brand-name drug products.” Such a classification is neither logical nor definitive. A useful classification, permitting meaningful comparisons, identifies each drug entity as containing single-source or multiple-source drug products. Single-source drug products are so classified because they are.