We use your email only to reply to you.
arXiv · 24 March 2016
By Nassim Nicholas Taleb
P-values follow a highly skewed meta-distribution, so the minimum over a few tests can sit far below the true p-value and results replicate poorly.
No summary of this paper yet.