Hacker News new | past | comments | ask | show | jobs | submit login

I forget the design of the experiment but I remember feeling that my prior assumptions (which were in line with GP) were potentially wrong, so it must've been moderately convincing. I work in clinical trials so I'm not a complete buffoon on experiment design, but accordingly I'm also aware a good experiment is obscenely difficult to conduct, and obviously this was nothing close to an actual RCT.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: