Sibylla Bostoniensis

But the scientific method is a method of testing hypotheses, not generating them. Science, reasonably enough, can only comment on the hypotheses submitted to it.

This is HIGH on my list of "things I wish everybody understood."

In my context (agriculture/rural int'l development), this is linked to the (bad) idea that "more math" = "more scientific" and also that "more scientific" = "better" so all the social scientists keep trying to demonstrate how scientific they are by doing more math. When really we'd all be better off if they could e.g. do good ethnography when that's the best way to answer the question at hand, rather than demanding that everyone do surveys and quantitative data analysis because that's "more scientific" even if it doesn't actually tell us anything.

....descending from my hobbyhorse...

It seems like in some ways that's paralleled in the "psychiatrists vs therapists" dichotomy you mention (in...someone's comment): because psychiatry is "more scientific" it's better, and the way to demonstrate that it's better is to keep getting more and more into (bio/physical/chemical) neuroscience. And when you add that to everything else you've talked about here, even that neuroscience is reduced to nonsense. Because while the scientific method is very good at answering specific, bounded questions, it's crippled without other ("non-scientific") ways of figuring out what questions to ask (and how to ask them).