### Yoav Benjamini - Challenges in Large Selective Inference Problems

183
views
0

0
9 months ago by
0
9 months ago by
It's interesting that, as you mention, statisticians recommend a new decision rule $p \leq 0.005$. It reminds me of the high-energy physicists insisting on using a $5\sigma$ rule for claiming discovery. Here is one paper by Louis Lyons https://arxiv.org/abs/1310.1284

and another:
https://arxiv.org/abs/0811.1663
0
9 months ago by
In case anyone is interested, Stephen Walker & I wrote a short note with a different take on the p-value controversy, forthcoming in American Statistician. Here's a preprint: Click Here