Scandinavian readers might want to take a glance at these posts on the Ekonomistas blog where Robert ÷stling address some common problems in a lot of social science.
As a political scientist, I’d say that:
ad 1. Yes, you’re right: Political science question the basis of actors preferences. Economics could be said to be a the same time deeper and more limited in its analysis.
ad 2. Oh, dear. You could waste a lot of time in methods classes on this. But my hunch always was that (e.g.) structural theories of discrimination have a tendency to create a closed room from which there is no escape. Mark Lilla, in The Reckless Mind, has some interesting things to say about radical theorists.
ad 3. In a way, I use the opposite strategy. Statistics are great when I want to get an overview of an issue. This then is a useful starting point for more detailed analysis.
ad 4. Non-economists would say that economics is method- rather than problem-driven. This may be an issue when academics and politicians want to apply economics to the real world.