"It's easy to get a false positive risk of 40% withOUT any cheating at all"
Reposted from
David Colquhoun
That's really good. I didn't show enough skepticism about p values when writing in the late 1960. But from a Bayesian point of view, it's easy to get a false positive risk of 40% with any cheating at all (when testing a point null). www.tandfonline.com/doi/full/10....
Comments