You can't answer "Have any famous tests replicated?" with "out of 100 studies from the year 2008, 36% replicated" unless one of those studies was actually famous.
Only if you're a frequentist. A Bayesian would see evidence that studies in general fail to replicate, and thus have a better prior for famous ones than 50:50.
Noumenon72|1 year ago
acover|1 year ago
exe34|1 year ago
Only if you're a frequentist. A Bayesian would see evidence that studies in general fail to replicate, and thus have a better prior for famous ones than 50:50.