3 ms·
No. P values this low are very rare. I haven't checked any of the analysis with my own data, but the approach seems pretty legitimate. Each is calculated in
by londons_explore 5y ago
No. P values this low are very rare. I haven't checked any of the analysis with my own data, but the approach seems pretty legitimate.
Each is calculated in a different way, using mostly-non-overlapping evidence, so it's almost legitimate to multiply them together to get P<1.3x10^-18... Not that there's any need - each alone is very convincing.
The whole lot really should have been written as three separate papers, with better written abstracts, and a metapaper, written by different authors, combining the evidence and comparing to the evidence for other hypothesis.