Post #3022641
2026-04-16 09:46 UTC
@jtonline@mastodon.me.uk
Personally, I would record the fact that 63 of my 263 survey respondents answered a version of my survey that didn't have the questions about THE PRE SPECIFIED PRIMARY OUTCOME in the study participant flowchart (not just the narrative of the methods section)
But frankly I wouldn't consider a 263-participant analysis to be "a large study" either so maybe it doesn't matter for the purpose of evaluating the blog post
Replies (1)
-
@nyhan@fediscience.org 2026-04-16 09:54
Actually it is worse than I noticed the first time I read the paper. I'm not really clear how they got down to 186 respondents with burnout info (the PRIMARY OUTCOME I remind myself). The difference between the 200 number that I expected and the 186 whose data is included is bigger than the number of respondents whose data is excluded because they did <5 consultations with the ambient scribe. Did some respondents skip those questions maybe? Too bad the participant flowchart doesn't say. But at least this statement is pretty clear: "Among 186 participants included in the burnout models,"