Small-scale Inference: Empirical Bayes and Confidence Methods for as Few as a Single Comparison

  • Empirical Bayes for One or More Comparisons
  • David R. Bickel
  • International Statistical Review, August 2014, Wiley
  • DOI: 10.1111/insr.12064

How to estimate the local false discovery rates of one or more hypothesis tests

What is it about?

This paper applies methods of estimating the local false discovery rate to one or more hypothesis tests.

Why is it important?

In testing many hypotheses, the importance of local discovery rates is that they strike a balance between the excessive false negatives of adjusted p values and the excessive false positives of non-local false discovery rates. This paper explores the possibility that the advantages of the local false discovery rate also apply to the case of very few hypotheses or even to one hypothesis.


David R. Bickel
University of Ottawa

This is a first step toward modifying estimators of the local false discovery rate for the very smallest numbers of hypothesis tests.

Read Publication

The following have contributed to this page: David R. Bickel