We have some mixed polling news today as we close out another week. The numbers are below, followed by my analysis.
Romney 49%, Obama 48% -- Romney +5% since 3 months ago (CNN)
Romney 48%, Obama 45% -- Romney +8% in since 1 month ago (FOX News)
Romney 51%, Obama 46% -- Romney +1% since 1 week ago (Rasmussen)
STATE AVG: Romney +2.5%
This is a pretty good batch of numbers for Romney overall, particularly the Public Policy polls of Iowa and New Hampshire. The Rasmussen numbers in Florida, Virginia, and Wisconsin are largely unchanged from a week ago, a good sign for Romney that he is holding on to his gains following what most consider a debate loss for him earlier this week. The CNN Florida number is also a good one for Romney. Taken as a whole, Florida has shifted dramatically to Romney in the past 2-3 weeks, but the likely voter screens continue to show a giant gap from registered voters, in the range of 5-7% in most polls. Obama's lead in Ohio in the FOX News poll, while smaller than it was pre-debates, is a good sign for him that his #1 firewall state remains in his column. The state polls, so far at least, are not showing any real signs of an Obama post-debate bounce. We started to see Obama's numbers stabilize a bit and Romney's bounce fade slightly at the end of last week, and that seems to be the same place we are at this week even after the second debate. I will add my usual caveat on state polls: stick to the polling averages, don't give too much importance to single, individual poll numbers, particularly from robo-pollsters like Public Policy and Rasmussen. If you must give credence to individual pollsters, stick with the high-quality ones that use live-interviews and traditional methodology -- NBC/WSJ/Marist, CBS/NYT/Quinnipiac, ABC/Washington Post, CNN-ORC, FOX News, etc.
It's also worth noting how big the registered voter vs. likely voter gap continues to be in some polls, particularly CNN and Gallup. In that CNN Florida poll, Obama actually is up by 7% among registered voters, but down 1% among likely voters. Showing a 5-8% pro-Romney swing when switching from registered voter models to likely voter models is somewhat of a head-scratcher since the usual historical gap is in the range of 2-3%, not 5-8%. The FOX News Florida poll, for example, has a more logical gap -- Obama trails by 3% among likely voters and by 2% among registered voters, a 1% swing. In Ohio, the FOX News poll shows Obama up by 3% among likely voters and by 5% among registered voters, a 2% swing. If pollsters like CNN and Gallup are over-exaggerating the 5-8% swing they're showing by even 1-2%, it changes the state of the race completely in every swing state. Florida goes from lean Romney to toss-up, Ohio goes from lean Obama to fairly safe Obama, etc. I don't want to get in the habit of nitpicking polls, but it's worth noting and will be interesting to see how things turn out on election day and whose likely voter model ends up being vindicated -- the ones showing a giant 5-8% gap or the ones showing a 2-3% gap. It's also another reminder of how turnout and ground games will decide this election. If Obama gets more of these registered voters that pollsters are wiping out in their likely voter screens to vote than expected, it will lead to him over-performing polls.
The national numbers are better for Obama today, as he is now tied or leading in every national poll/tracker except for Gallup and has regained a very slight lead in the national poll average for the first time in two weeks.
* RAND: Obama +3% (49%-46%)
* Reuters: Obama +3% (46%-43%)
* UConn: Obama +3% (48%-45%)
* IBD/TIPP: Obama +2% (47%-45%)
* Public Policy: Obama +1% (48%-47%)
* Rasmussen: TIE (48%-48%)
* Gallup: Romney +6% (51%-45%)
Pretty clear there which poll is the outlier. Obama has gained about 1% in the average of national polls this week. Some will view that as a bounce from the debate, others will view it as the continued stabilizing and slight fade for Romney from his peak that began last week.
The 538 forecast has Obama back up to a 70.4% chance of winning, his highest point since October 9th when he was in the middle of his post-debate free-fall that took him all the way down from 87.1% to 61.1%, so he's recovered about 9% of what he lost.
As I predicted back in my debate review, Romney has held on to a lot of the gains he got from his first debate win despite losing the second. Some permanent changes were made in that debate in the eyes of many voters -- Romney went from being the monster on the 47% video and the caricature the Obama campaign had turned him into to an "acceptable alternative" to some swing voters, particularly right-leaning Independents. The value of that is immeasurable and once that change happens, it's usually hard to undo. The electorate seems to have hardened a bit as we hit the final stretch. That pretty much ensures us that we are headed for a close finish in this election, similar to 2004.
Where we are at now is not very different from where we were at prior to the conventions -- a very close race, with Obama having a stubborn, persistent edge in the electoral college. Assuming the likely voter models that polls are using, which wipes out about 5-7% of Obama's margin in some polls when compared to the registered voter models, is accurate in its extremeness, the Ohio-Iowa-Wisconsin-Nevada firewall is the main thing holding Obama up right now (even if you take Iowa out of that, just the Ohio/Wisconsin/Nevada combo is enough to get Obama to 271 and victory), and those are the states worth keeping the closest eye on. Florida and North Carolina both continue to look favorable to Romney, particularly North Carolina which I've been saying for months is a waste of resources for the Obama campaign in this environment (if they win it, it will be part of a larger electoral landslide, not the nail-biter we are headed for -- North Carolina is not a tipping point state in any realistic scenario and so it's hard to justify spending money there that could be spent on further solidifying states that could actually make the difference between winning and losing like Ohio, Wisconsin, Nevada, and Iowa). Virginia, Colorado, and New Hampshire look like the true toss-ups right now, with Ohio, Iowa, Wisconsin, and Nevada all leaning blue and Florida and North Carolina leaning red.
Be sure to follow me on Twitter for more of my election analysis.