Partly this is because the race on the Democratic side looks pretty close, and a close race is unlikely to provide significant momentum for anyone. But here's the main reason: in the past, except for 2004, Iowa hasn't been all that important. Gore was way ahead of Bradley in 2000 regardless of Iowa; Clinton came in fourth in 1992; Dukakis third in 1988; and Mondale's 1984 win was routine. For my money, you have to go back to 1976 to find another year in which the Iowa caucus really made a big difference.
But we've all been shell shocked by 2004, when John Kerry seemingly came out of nowhere to beat Howard Dean and never looked back. It made Iowa look superhuman, and I have a feeling that Democrats have never gotten over it. So for a long time everyone has been expecting 2008 to be a repeat of 2004, despite the fact that 2004 was sort of a weird one-off fluke.
At least, that's my guess. Obviously early states have more influence than later states, but I really do think 2004 was a fluke. Iowa this year has the potential to end the John Edwards campaign, I think, but I doubt that it has the potential to crown a winner. We'll have to wait for February 5th for that.
And on the Republican side? My prediction is for chaos. And I'm looking forward to it.