I don't think economics really exists as a separate discipline. I think it exists as a kind of academic cult cursed by delusions of infallibility, but the practical policy foundations are always based on politics, sociology and psychology - with a bit of game theory and statistical analysis at the more rigorous end.
In terms of real-world decision making, I think it really isn't possible to understand economic relationships except in terms of politics, sociology and psychology. They're not footnotes or distractions, they're fundamentals.
Politics obviously plays an enormous role, because governments are enormous actors in the economic arena. Psychology's role is growing -- there is actually a sub-discipline called Economic Psychology, although it's still in its infancy -- because psychologists can help economists develop a stronger understanding of (say) happiness or buying habits or or or.... Sociology and economics have always had a great deal of overlap, and academics from each work together quite often. (My uncle, who is a retired sociologist, is an example.) If I wanted to write a book on Africa's current economic situation, I would likely need to get in touch with someone who understands African culture and politics, because it would shed a great deal more light on the issue for me. Proposed solutions to problems are, obviously, useless if they do not work properly with the cultural and political situation (as when FDR had difficulty pushing Keynesian worker programs in an era of American politicians seeking a consistently-balanced budget).
As far as infallibility is concerned, in any field -- in my experience, at least, and I must admit to having little in the hard sciences (so I'll leave that to Miguel, Colman and anyone else who has said experience) -- you'll find people who take this view of their discipline, but I think you'll find plenty of examples, especially among the big names, of economists who do not believe in the infallibility of economics. In fact, the book Miguel wrote this diary on was one built to challenge a previous failure in the field, as were Friedman's writings on things like the Phillips Curve. Conservatives want live babies so they can raise them to be dead soldiers. - George Carlin