Hacker News Reader

Be careful interpreting this study summary.

The article's peer reviews (on PubPeer) list a few analytic issues that call the findings into question:

https://pubpeer.com/publications/59F515775D10D854ABB7F4B31D4...

Hrmmmmm the controls for the first 2 graphs appear to be the same and they're called out for it [0].

But it could very well be that they're supposed to be the same? Authors may just need to better highlight why the same control is allowed to be use for comparing 2 different result sets.

[0] https://pubpeer.com/publications/59F515775D10D854ABB7F4B31D4...

Yep, but other issues are discussed too. Suspicion is warranted, though signs of errors and fraud require even deeper investigation.

Reply on news.ycombinator.com