{"id":"arxiv/2603.08002","url":"https://arxiv.org/abs/2603.08002","tags":["math.ST","stat.ME"],"score":0,"title":"Post-Hoc Large-Sample Statistical Inference","domain":"math.ST","source":"arxiv","authors":["Ben Chugg","Etienne Gauthier","Michael I. Jordan","Aaditya Ramdas","Ian Waudby-Smith"],"reviews":0,"subject":"math.ST","summary":"We derive inferential procedures for large sample sizes that remain valid under data-dependent significance levels (so-called \"post-hoc valid inference\"). Classical statistical tools require that the significance level -- the \"type-I error\" -- is selected prior to seeing or analyzing any data. This restriction leads to some drawbacks. For instance, if an analyst generates an inconclusive confidence interval, repeating the process with a larger significance level is not an option -- the result is final. Recently, e-values have emerged as the solution to this problem, being both necessary and…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-09","reactions":{}}