The problem is that a reproduction crisis, on its own, doesn't mean that a field is doing a bad job. With systems that involve lots of moving parts, the prior probability of finding a correct model is just dismally low. Bayes simply guarantees that the vast majority of our positives will be false in this case. We probably need to utilize more sophisticated statistics in this case in order to capture a more nuanced notion of a study's utility.

Worse, most of the time I hear people criticize psychology, the person is implicitly using this as an excuse to not critically examine their personal beliefs about how human minds work.

Paychology is hard, and we're feelinng around in tye dark, but chemistry had the same problem 300 years ago with ideas like phlogiston theory.