Rejoinder on: Control of the false discovery rate under dependence using the bootstrap and subsampling

摘要

We are extremely appreciative of the insightful comments made by all the responders. The goal of constructing useful multiple testing methods which control the false discovery rate and other measures of error is currently a thriving and important area of research. On the one hand, the bootstrap method presented in the present work seems to work quite well and is supported by some theoretical analysis. On the other hand, many more important practical, computational, and mathematical questions remain, some of which are addressed by the responders and which we touch upon below. We also appreciate the added references, which help to provide a more thorough discussion of the available methods. Our paper was the development of a particular methodology and was by no means a comprehensive account of the burgeoning FDR literature.