Paul Krugman, Mark Thoma, Tyler Cowen, and Arnold Kling discuss ideological divides in the economics profession. Generally, libertarians are expressing skepticism about economic stimulus, while liberals are wholeheartedly supporting it.
Though some are alarmed by this, it shouldn't be surprising. Changing one's mind is not an effortless process, so people will try to avoid it. In this case, economists approach the issue with a bias in one direction, and do the best they can to rationalize it.
It's especially noticeable in this case, because
1) There aren't strong incentives to be correct in this situation. A single economist is unlikely to have a large effect on the stimulus, and offering an incorrect forecast every once in a while won't destroy anyone's reputation. If, say, one of these economists was personally in charge of administering the stimulus, the bias would probably shrink. This fits with Bryan Caplan's model of rational irrationality. (Caplan himself is a blogger on EconLog with Arnold Kling.)
2) There is room for disagreement. As Arnold Kling notes, "There are no controlled experiments in macro." Thus, rationalizing is easy.
Prediction markets, like Intrade, largely solve the incentive problem, so I prefer them to public debate. It would be fascinating to see, for example, Paul Krugman and Tyler Cowen make a bet on this. However, prediction markets can't do the impossible-- they can't predict the future if the necessary information just isn't there.