A lot of macroeconomics in particular seems to follow aristotlian thinking, i.e. you can just logic your way from step to step. It is interesting in that pure mathematics does work this way in that you can start with a bunch of axioms/definitions and combine them and work with them until you reach a conclusion without any need for experimentation. For example, you don't need to measure a bunch of triangles to show that the Pythagorean theorem is true.
The way i was taught macroeconomics in school, it takes the mathematical approach of starting with a list of assumptions that seem to be true, and using mathematical logic to derive conclusions. The trouble is that: 1) many starting assumptions aren't as true as you'd think, and 2) many of the "logical" steps also aren't as logical as you'd think.
Supply and demand is an easy example. The classic idea is that as supply goes up, price goes down, and vice versa, while as price goes up, demand goes down. The existence of Veblen goods (i.e., things that people want because they are expensive) shows that the demand part isn't right.
None of this would be a problem if it was just an intellectual exercise to help develop hypotheses for use with more sound scientific methods, but it's used for policy directly.