It’s not respected research, though, is it?
At least 3 of us have poked massive holes in the methodology. I can think of at least two ways of restructuring the analysis to make it more robust.
I’m very disappointed this was published in PNAS, that’s usually a quality journal. I had a paper published in PNAS 10 years ago so I’m bound to say that. We had much more rigorous statistics than this paper.
If I’d have been reviewing it, I wouldn’t have let it pass. However, I know how the academic publishing racket works, so I’m not surprised.