search

aaronpeikert / bayes-prereg

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

be more forcefull about type I not equal theorethical risk #90

Open aaronpeikert opened 10 months ago

aaronpeikert commented 10 months ago

preregistration reduces uncertainty about the Type 1 error rate, and under the assumption that it is possible to inflate the Type 1 error rate when a study is not preregistered, reducing the uncertainty through preregistration increases the posterior probability of the hypothesis after a test that yields a significant result.