168: Meta-meta-science
Everything Hertz
00:00
Unspecified Replication Rate Targets
I think it's very difficult to actually define what is too low, but surely when 50% of key cancer studies do not replicate, you have to go. If you talk about the end-get study, it was certainly less than 50%. Oh, what the hell? Yeah. There's only really heuristics that go into something like that. You have an expectation that's built out of how explicit or exactly can a quarter method be written down in the first place? So I don't know. That one feels a bit soft.
Play episode from 28:05
Transcript
