Among the important gains in health care in this quarter-century has been the progress of disciplined statistical methods for clinical research. Thanks to leaders like Sir Richard Doll, J. N. Morris, Kerr White, Alvan Feinstein, Frederick Mosteller, Thomas Chalmers, David Sackett, and others, clinical investigators today use more sophisticated approaches to inference, and readers of their reports are more able to judge the soundness of experimental design. The randomized clinical trial (RCT) has emerged for good reason as the prince of designs, and we also understand better the proper use of case-control, cohort, and quasi-experimental studies. See also p 841. These new standards have reached the mainstream. Journal reviewers demand evidence of proper control and statistical analyses. Study sections attend carefully to details of proposed methods, like sample size calculations, randomization rules, and investigator blinding. Of the first 50 original scientific reports in the New England Journal of Medicine in