Discover and read the best of Twitter Threads about #actstat
Most recents (1)
Join us at #ACTStat Baltimore. I look forward to discussing methods for #perprotocol effects in randomized #trials bit.ly/2uPlGpZ
Because, remember, intention-to-treat analyses may not be conservative when comparing two active treatments… bit.ly/2uZP1hS 
… and, in fact, intention-to-treat analyses may not be conservative even in #trials that compare active treatment vs. placebo
