Atheists and suicide

[Angrboda]

Beyond the problems of methodology, these approaches are testing the hypothesis that there is something about the life experience of the religious which makes them less prone to suicide. Now you can talk about anomie and all sorts of adaptationist "just-so" stories, but until you demonstrate this difference, quantify it, and show that it varies in proportion to suicide rates, you're asserting an effect without a plausible mechanism (plausible in the sense that science uses the term in reference to hypotheses). If you don't have a mechanism, and in this case, the mechanism is largely assumed and assumed to be operative in the religious as a whole, and not in the non-religious, then statistically, it is legitimate to push the bound for Type I errors and require greater statistical significance. (This is colloquially referred to as "extraordinary claims require extraordinary evidence". Victor Stenger notes that while medical studies require a 0.05 significance, the standard in physics is an order of magnitude higher, and he points out that this has the practical effect of a large proportion of medical studies giving a positive result, only to be overturned later.) So even if there is a weak statistical effect there, that isn't sufficiently significant to be taken as real.