3 ms·
the previously accepted statistical threshold of "scientific truth" This concept heartily amuses me. Care to say more?
by philwelch 16y ago
the previously accepted statistical threshold of "scientific truth"
This concept heartily amuses me. Care to say more?
- hga 16y agoI think it was a meta-analysis (a study using multiple original studies) and I recall 95% being a number of significance. As in it was either the previous or new, improved and of course lower threshold of "truth" (or rather statistical significance or whatever). I'm going by a 2nd or 3rd hand report, but the reported details are simple and easily falsifiable. (ADDED: Or maybe it's 90% and they used 85%??? Anyway, that should give you enough to search on if you want to follow it up.) Epidemiology is hard. When it's about a politicized topic (and almost all are nowadays) I've learned to just not pay much attention to any one study, at least until others have a chance to tear it apart.