Is it any better in fields other than math and the natural sciences? For medicine and nutrition we're constantly backtracking on our previous discoveries. First it was fat that's bad, then it was sugar.
Nutrition is really complex and hard, and we learn more every day. But for the most part, it's not like complete bullshit is invented whole-cloth on a piece of napkin. It slowly builds and expands on prior knowledge; as far as I know it's not really "constantly backtracking".
I’m not persuaded that economics has that quality. It seems to be infected by political interests too easily.
As evidence of that statement i submit the example of Austerity as a concept. It’s something that’s been applied hundreds of times, there was a resurgence in the 2010s but it was very popular after the First World War too.
One problem, despite all these applications of the concept, its only ever worked twice in documented history. Both instances occurred in a context that doesn’t apply today and, well it’s not impossible but almost impossible for that context to appear again.
The concept has no economic foundation, it’s a device born of a political belief. It even contradicts economic assessment.
And yet, It’s continually proposed. It’d be like suggesting sugar is a weightloss product if you consume enough. If you gained weight, you didn’t do it right, you just need to consume more sugar, then you’ll lose weight.
This is the case for some other social sciences as well, but I have the impression it's (much) worse for economics. The effects are certainly a lot more impactful.