Low Power and Striking Results — A Surprise but Not a Paradox

Publisher Website
Google Scholar
Abstract
In this issue of the Journal, Verdonck et al.1 report on a study that, as they planned it, was barely large enough to detect a massive advantage for a controversial and expensive therapy. Moreover, the way they calculated the size of their sample was too forgiving: it presupposed a one-sided test of significance, which ordinarily would not be acceptable in these pages. Yet the study does appear here, and it provides evidence strong enough to guide therapeutic decisions. What went right?Verdonck et al. studied patients with non-Hodgkin's lymphoma who had partial responses to three cycles of CHOP therapy (cyclophosphamide, . . .

This publication has 2 references indexed in Scilit: