Work

After the data choose the report: what multiplicity, selective, sequential and post hoc guarantees still support

Year: 2026
Type: preprint
Abstract: Statistical reports should be traceable to the guarantees that support them. Multiple-testing, selective and sequential guarantees are stated for a procedure applied as planned, yet analysts choose after the analysis what to report: part of a list, the results that were selected, the moment to stop, or the level at which a result counts as significant. This review argues that a guarantee does not carry over to such a choice by its name: it supports a specific sentence about the procedure in repeated use, and that sentence must be re-derived after each choice. For each operation, the review states the sentence each guarantee still supports, the assumptions it then needs, what is lost, with an exact small counterexample, and what the alternatives, including e-value and post hoc guarantees, control instead. For example, control of the false discovery rate (FDR) at level q supports a statement about the average proportion of false discoveries over repeated analyses, not that at most a fraction q of the reported discoveries are false, and online FDR control at fixed times does not carry over to a time chosen from the data. Each relation is credited to its source or marked as derived here. The supplement and the exact-computation code (https://doi.org/10.5281/zenodo.24127934) accompany the article.
Language: English
Open Access status: green
OpenAlex ID: W7234418710
Cites:
Cited by:
Related to: 10