ABSTRACT
Pocock presented a group sequential design, which has the format of a k-stage test. He presents exact results for two populations which are normally distributed with known variance. Chase and Hewett discussed a method of constructing two-stage tests that only requires that the population(s) be of the continuous type and p-values can be computed. The procedure is based on p-values and results from a modification of Fisher’s method of combining independent tests. Simulation results indicate that the tests maintain good power and achieve savings in expected sample size. Madsen presented a k-stage test procedure using p-values. This procedure is distribution-free in the sense that it is applicable in any situation where a p-value can be calculated.
