Two-Stage Sample Size Prediction Methods Applied to Clinical Trials

Various two-stage sample size prediction methods have been proposed in the literature for comparing two means. These methods can be viewed as falling into a family of two-stage methods that provide exact solutions to the Behrens-Fisher problem. The general family of methods is outlined and the special cases that appear in the literature are reviewed and extended to a more general collection of solutions. Distributional results are given for testing if there is a difference in the means of an intervention and a control in a clinical trial.