Several recent publications have identified important methodologic problems in the design and conduct of antimicrobial trials in surgery. Developed by consensus of the members of the Scientific Studies Committee of the Surgical Infection Society, this report provides broad guidelines for the construction of antimicrobial trials. The central issues identified include pretrial definition of study purpose, entry criteria, assignment device, and statistical analysis. These issues are fundamental in designing studies with an acceptable likelihood of finding differences among those antimicrobial regimens at least risk to the study subjects. The importance of stratifying patients on the basis of background condition, disease, and severity of illness is stressed. The inclusion in a study of variables that enhance the statistical power and, therefore, the believability of this study is stressed as an important means of clarifying substantial differences between therapies.