3 ms·
There are very good reasons to do exactly that. I don't know how often you work with complex models like this - but I do all the time. They are very often wrong
by darawk 6y ago
There are very good reasons to do exactly that. I don't know how often you work with complex models like this - but I do all the time. They are very often wrong for very subtle reasons. It's extremely difficult to satisfactorily prove something in a descriptive analysis, especially of a complex, dynamic system.
By default, you should basically never trust any descriptive analysis of a system like climate or the economy. Climate change specifically has received so much scrutiny by now that it's probably trustworthy. But in general you should have a very low prior on this entire category of research.
- delecti 6y agoYou seem to be deliberately avoiding my point. The fact that some complicated models are wrong is irrelevant. The fact that you've worked with lots of complicated wrong models is irrelevant. The rational default assumption for a lay person, when the experts say that their complicated model agrees with the simplified model which can be explained to the lay person, is to tend to accept what those experts say. It makes no sense to go from "I understand the grade school science behind climate change" to "but those experts use complicated science, so they're probably wrong".
- darawk 6y agoYes, the rational default for a lay person is to accept expert consensus. But the question at hand is what weight they should give that consensus. And that weighting should change, depending on the nature of the system being understood. We're talking about credences, not point estimates of beliefs.