As is so often the case, my colleague Bryan Caplan writes  with far greater concision, elegance, accuracy, and insight than I ever manage to muster when I try my hand at such explanations (for example, here  and here ). Here are the opening two paragraphs of Bryan’s superb post:
What is economic theory? Is it a body of proven truths? Or a set of hypotheses whose only merit is that they’ve so far been successfully tested against the facts?
Most economists openly embrace the latter position, but secretly believe the former. The smoking gun: The typical economist has well-defined views on a wide range of issues, even though he is only familiar with a handful of empirical literatures. If economists really believed their official methodological position, they’d be agnostic on the vast majority of topics. They aren’t.
Bryan’s post reminds me, in spirit if not in detail, of Ronald Coase ‘s insightfully expressed skepticism – in Coase’s 1982 article “How Should Economists Choose?” – of Milton Friedman’s defense of positive economics. (I can’t find a link to Coase’s article, but David Henderson, back in September, posted a nice quotation from it .)