Friend reminded me of this paper by a CMU grad and professor. Good point well made; I'm reminded of what a methodological wild west it is on frontlines of science. Occasionally scientists publish OP-Eds valorising this with nerd machismo. I'm unimpressed. https://www.andrew.cmu.edu/user/ddanks/papers/BayesCommit-PubVersion.pdf …
-
Haven't you previously described this as "methodological pluralism" and lauded it?
Hmm, so I think of the practice in this paper that its not that they have some unusual method which legitimises their inferences, but they are just failing to do as they intend. But I can't say why I think that's bad while pluralism is good! I guess I'm just not quite Feyerabend?