search

aaronpeikert / bayes-prereg

https://aaronpeikert.github.io/bayes-prereg/#2
Creative Commons Zero v1.0 Universal
6 stars 6 forks source link

implications of using probability vs credibility #54

Open aaronpeikert opened 1 year ago

aaronpeikert commented 1 year ago

Andreas:

Personally, I think that a lot of confusion with Bayesian results is that the 'term' probability makes it sound objective and truthful; would you be okay with replacing probability with credibility to underline the subjective nature of this approach? Or would this be even more confusing?

aaronpeikert commented 1 year ago

@brandmaier I am against using credibility. Probability is a mathematical construct. We use it strictly as a mathematical construct when employing it in Bayse's theorem. Only when we assign the posterior the term "epistemic value" do we move away from mathematics to epistemological. However, we could make the same arguments for all other measures of confirmation/corroboration that monotonically increase with increases in posterior. That is, to my knowledge, all formalized measures.