What a poll tells you is rooted in the methodology. As I noted a few posts back, if you use the right methodology you can get a Romney victory in the 64-36 range. Were my assumptions warranted? Maybe. They represent my opinion and it's as good as anyone elses, so there.
Here is an article showing the basis for some of the results we've been seeing, and here is the graph:
Note Rassmussin, the lonely outlier, giving the R's a 2 point advantage. This would account for the 2010 election, the continued enthusiasm on the right, and a distinct lack of it on the left.
I'm sticking with my original call: Romney by 5, except in Illinois where with 110% voter turnout, Obama ekes out a razor-thin win.