Hacker News Reader

You're assuming also that peer review will fix statistics issues. Statistics will often fall through the cracks.

That's why I mentioned meta-analyses as being the most important in the end. No one paper in these fields is enough to make any conclusions.