Don't like p-values? We're presenting a completely new interpretation in the context of clinical trials. With Andrew Gelman, Sander Greenland, @guido_imbens Simon Schwab and Steve Goodman. @Lester_Domes @goodmanmetrics @NEJMEvidence

RT @vientsek: Chat with Bayesian statisticians in your local area NOW and **DOUBLE** your RCT's effective sample size by using this ONE WEI…

Great to collaborate with Yefeng Yang, @itchyshin and @nikosIgnatiadis to study the signal-to-noise ratio in eco evo studies, and what we can realistically expect from replication studies.

@statsepi You're missing out: Data were presented as mean with standard deviation (SD) when normally distributed. Skewed data were presented as median with interquartile range (IQR). Categorical data were presented as number with percentages. P-values < 0.05 were considered to be signif.

RT @Collaboratory1: 📣 Webinar recording and slides now available: "A New Look at P Values for Randomized Clinical Trials" with @ErikVanZw…

@drjohnm @Sensible__Med Thanks for having me on the show, and for taking a real interest in statistics!

@f2harrell @drjohnm @Sensible__Med Thanks for commenting, Frank! I tried to answer some of your remarks on @Sensible__Med. If you have time, please consider joining my NIH talk tomorrow (April 5). I can explain in more detail what we did and why, and we can discuss some more!

RT @drjohnm: I had the pleasure of talking with @ErikVanZwet this morning — podcast on p-values coming soon to @Sensible__Med I’m sure th…

@Collaboratory1 @bds_lumc I'll talk about "A New Look at P Values for Randomized Clinical Trials"

RT @Collaboratory1: In this Friday's PCT Grand Rounds, we welcome @ErikVanZwet of @bds_lumc at Leiden University Medical Center, who offers…

@HenryWyneken2 @f2harrell @lakens @david_colquhoun I would like to understand what severity is. Perhaps you can help me? Suppose z1 and z2 are independent N(mu,1). I _think_ the severity function in that case is SEV(m,z) = P(z2 > z1 | mu=m, z1=z) Is that right?

@PloederlM @ueafam @guido_imbens @Lester_Domes @goodmanmetrics @NEJMEvidence (2/2) In the NEJM/E paper, we treat the CDSR as a large collection of trials. In fact, it's a collection of meta-analyses. That makes it possible to study publication bias (e.g Hedges, 1996). It's my preliminary impression that there is some publication bias but not very much.

RT @NEJMEvidence: Original Article: “A New Look at P Values for Randomized Clinical Trials” by @ErikVanZwet, PhD, et al.

@MTimC1 @pash22 Yes, this is the correct plot. And I believe this is a toll free link:

RT @pash22: A New Look at P Values for Randomized Clinical Trials via @ErikVanZwet et al

@EikoFried Thanks for posting! The toll free link is here: Please note that the labels in the Figure are reversed - hopefully to be corrected soon!

@damekdavis @beenwrekt That's not true. The NEJM Evidence paper has an online supplement with the code. Moreover, I've just provided the data selection for an earlier paper (van Zwet, Schwab and Senn) in a comment on Ben's blog.

@gushamilton @beenwrekt @stephensenn @Lester_Domes Gus: Thanks for pointing this out to me. I've commented on the blog post.

@mihagazvoda @PhDemetri @Alan_Taylor_314 Where there's a will, there's often a way. Please feel free to email me if you'd like to discuss.