With apologies to those who downloaded the earlier version of the Replication Report, I made a stupid mistake calculating effect sizes. They are corrected, and a few implications are slightly altered, in the current version of the document available here. Thanks to Florian Cova for pointing out my mistake.