This is a great statistical analysis and it was a pleasure to read, but I wasn't expecting the claims to be so poorly supported. There's also a reply from one of the Meta authors there, worth checking out.