1/ A new "Masks RCT" has been doing the rounds lately with "The miracle of masks". Don't say more. What do you know when your CI cannot rule out no-effect or harm? Yeah, that's correct. You drop it in the garbage can where it belongs. 
https://twitter.com/beenwrekt/status/1433117497725624324
2/ I wouldn't even going to comment on it, a cursory look at the tables already show that whatever math you apply to it was going to be anything... but given that I got bombarded with "What do you think about the masks RCT" it forced me to comment.
3/ Looking at the tables is a good time saver. I just didn't have to read 90 pages where a single table would suffice to tell me everything I need to know.
4/ I will leave you with my usual phrase when this happens repeatedly: "If you have to work THAT hard to find a signal (any signal), the most likely case is that there is none".
5/ Just to explain what I mean with the first tweet. I am all for publishing experiments that fail. But they should be reported appropriately as such. The Danmask study been to me probably the best example of it. I am very wary with studies where the crude and adjusted change.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
