Ras national poll has a Dem +2.5% to 4% turnout model over Rs for 2012.
I think it will be at least even (like 2004) to R +2%.
Rasmussen last self partisan ID of R +4.3% (37.6%) over Dems (33.3%) with a sample size of 15,000.
Rasmussen ignores his month to month partisan poll because he has already set his baseline for this election.
Rasmussen kept using a D+2 assumption even after its August polls showed likely voters to be R+4, probably just to play it safe, since D+2 is closer to the historical norm (it was D+3 or so in 2000, even in 2004, and D+7 in 2008). It is possible that his September party-ID numbers (which he will release this coming week) will not be as Republican, since the RNC was in August and the DNC was in September. If the electorate on Election Day ends up being D+2 or less, Romney will probably win; if it’s D+3, it will be very close, à la 2000, and it could go either way; and if it’s D+4 or more, Obama will likely win.
What is clear is that the electorate will not be D+7 this year, much less the D+8 through D+12 that some ridiculous pollsters have been using.