Matters of judgment arising commonly in applications of statistical inference in clinical trials research include: the intent-to-treat analysis (treated briefly), effective use of multiple comparisons procedures, the nearly universal appropriateness of onesided hypothesis testing, and the extraction of treatment-by-stratum (in preference to treatment-by-center interaction) in the analysis of multicenter clinical trial data. This essay contributes basic guidelines resolving such issues
Food and Drug Administration.Guidelines for the format and content of the clinical and statistical sections of an application. Rockville, MD: FDA Center for Drug Evaluation and Research; July 1988.
2.
MillerRG. Simultaneous statistical inference (Second edition). New York: Springer-Verlag; 1981.
3.
HochbergYTamhaneAC. Multiple comparison procedures. New York: John Wiley & Sons; 1987.
4.
WilliamsDA. A test for differences between treatment means when several dose levels are compared with a zero dose control. Biometrics. 1971;27:103–117.
5.
FleissJL. Statistical methods for rates and proportions (Second edition). New York: John Wiley & Sons; 1981:260–280 Table A.3.
6.
FleissJL. Analysis of data from multiclinic trials. Controlled Clin Trials. 1986;7:267–275.
