Duncan Watts outlines it:
Politicians dealing with urban poverty feel they already have a good idea why people are poor. Marketers planning campaigns already feel that they have a decent sense of what consumers want and how to make them want more of it. Economic policy-makers believe that they can do a reasonable job of getting the incentives right, whatever end they are trying to achieve. They don't expect to get them right all the time, and they would be the first to admit that it's complicated. Nevertheless, they do think the problems they are contemplating are mostly within their grasp, that they are "not rocket science".
What's puzzling about this attitude is that if we compare recent progress in the physical versus the social sciences, it should be clear that we're actually much better at rocket science than managing the economy, merging corporations, or even predicting how many copies of a book will be sold.
The Dish 