Not bad news for Newt in any serious way. The RCP average probably is a better measure.
Most important:
Rasmussen has not revealed, or it has not been made public, how he is weighting his sampling of Iowa caucus goers.
If this were a national poll against a democratic opponent, we would be told how many democrats and how many republicans went into the result. We would know the books were cooked if a poll had 800 democrats and 200 republicans.
We are not told how many conservative, moderate, and liberal republicans go into Rasmussen’s numbers.
Since Iowa has been controlled by a very large group of conservative Christians for a number of cycles now, they should be the lion’s share of the sample.
We are not told and/or Rasmussen has not thought of it.
At this time in 2007, Rasmussen had Huckabee and Hillary leading their respective national polls.
He’s good, but not fallible, and Iowa is difficult to predict through polls.
That said, Rasmussen’s Iowa polls in December 2007 fairly accurately predicted Huckabee’s win and Romney’s second place finish there. It did overestimate Huckabee’s margin to some extent and underestimated McCain’s and Thompson’s performance.