If you: 1) Cite only the positive studies 2) Don't make any claims about effect size (i.e. correlation strength & predictive strength) You can basically make

If you: 1) Cite only the positive studies 2) Don't make any claims about effect size (i.e. correlation strength & predictive strength) You can basically make this point about literally any substance with enough studies on it. Ranging from melatonin to insulin, to idk... probably not mercury? So I guess that's a strike against 1 and 2 being fully generalizable.