Well said. It doesn't reveal problems with current methodology, it reveals problems with its application, and understanding. Larger sample sizes would fix quite a bit. Studies, even "pilot" or "priming" studies, are not reliable at a .05 significance. Especially when most of the false positives will get published, but nearly none of the false negatives will see the light of day.
