Comment by nerdponx

Yes, and it's a legitimate problem in science. In principle (not that anyone does this), you should probably apply some kind of FDR or FWER control procedure to the entire body of literature in each field. And that's before you get into the known effect of publication bias.