In terms of general e-processes, the work on testing exchangeability comes to mind. But is has been a long time since I looked into that and much has come out since then that I don't know of. Also, I still have to read Ryan Martin's paper to know what its flavor is ;).

Very interesting! Thanks for sharing!

Have a look at e-values:

en.wikipedia.org/wiki/E-values

Here is some intuition:
blog.vvsor.nl/2021/09/scie...

BTW, I can recommend to get familiar with this one:
www.medrxiv.org/content/10.1...
It helps to judge the trustworthiness of RCTs

While a lot of people might focus on how this got published in the first place (rightfully) I wonder if we shouldn't also consider this a pretty huge win for BMJ's data sharing policy.

Given increased use of single pivotal trials supporting FDA approval of new drugs & biologics, our study findings indicate a need for greater clarity & transparency by reviewers & sponsors on what evidence was considered to be confirmatory when approving drugs based on a single pivotal trial (7/8)

Radboud researchers René Aquarius @deadneanderthals.bsky.social and Kim Wever were about to embark on a literature review project when they noticed several odd things that made them stop in their tracks, then change the course of their study entirely.

Surprisingly, I4R has never been contacted by any of the authors' universities/funders. This is disappointing given that we documented irregularities in 10+ lengthy reports. That's easily 300 pages of data irregularities.

I like the plot. Don't see any error. Plot can be a visual aid to help the reader understand how regression to the mean might affect other things in the data that the paper is interpreting and how analytic choices are taking that into account.

An error might be in the analysis+interpretation!