Statistics Done Wrong: The Woefully Complete Guide

by Alex Reinhart

which makes p values “psychic”: two experiments with different designs can produce identical data but different p values because the unobserved data is different.