Professor Ioannidas, a Stanford epidemiologist and statistician (most famous for first calling attention to the problem with the standard use of p-values in medical research), obviously has a contrarian instinct, but sometimes those instincts are dangerous. I had been ignoring this piece, which seemed obviously unsound even to a layperson like me, but since it continues to circulate, I thought I would mention it and one of the decisive replies. (The comments on the Ioannidas piece also point out some of the obvious omissions from his analysis.) There is a longer interview with him here, where he sounds somewhat more reasonable.