November 6, 2012 - My Final Prediction
If you haven't voted yet, why are you reading this? Go vote!
I will not be doing any more poll averages after this, even with final polls that will drop today, like Fox. I'm going with the final polls that have been released so far, and making my predictions.
R+1 - Rasmussen Daily Track
R+1 - Gallup
O+1.6 - IBD/Tipp
O+3 - ABC/WaPo
R+1 - Battleground
O+3 - Pew
Even - CNN/ORG
O+0.005 - NBC/WSJ
Even - Monmouth
O+0.51% - Current RCP Average
O+1.40% - Average using the 2008 turnout model
R+1.52% - Average using the D+3 turnout model
R+3.85% - Average using the 2010 turnout model
R+3.92% - Average using the 2004 turnout model
R+7.02% - Average using the Rasmussen Party ID turnout model
Average Democrat advantage across all polls: D+5.00
Average Independent support for Romney across all polls: +8
I thought long and hard about including the CNN poll in the final analysis, and in the end decided I would do so. It heavily favors Romney, but I think it helps better capture the Independent support Romney has. For most of these polls I have posted in depth analysis, and I don't need to revisit those.
As I said before, I am offering three election scenarios. D+3, Even, and the Rasmussen Party ID.
D+3: (likelihood 10%) One of the reasons that I think the pundits and pollsters have been expecting a D+3 turnout this year is that looking at 2008, the Dems turned out at D+2, while the Republicans were at R-5 over 2004. I believe the thinking is that the Dems will still be able to turnout their base and execute their election strategy, but the Republicans will bring their base back to almost 2004 levels. I believe that this is the Romney campaign's turnout model. For weeks now, the only cold water that I have had thrown on my expectations has been from the campaign itself. I keep thinking this will be a blowout, then they leak some internals showing them barely ahead in Ohio. However, from their perspective this makes complete sense. They are still polling, and must apply a turnout models to determine the results. They have to assume that their competition will be able to execute their turnout strategy as well. You can't go into a turnout election assuming the Democrats won't turn out. Having said all of that, I think they have over estimated the Democrat's ability to turn out their base. The early voting numbers are down from 2008, and all of the campaigning on the other side has been designed to appeal to the base. Obama has made no effort to appeal to the Independents or undecideds. On the other hand, I think the GOP has underestimated the enthusiasm of their base. The size of the crowds at rallies has caught them by surprise, as has the number of volunteers.
Using this model Romney will win 49.21% to 47.69%
With this 1.52% win, Romney will take FL, VA, CO, NH, and OH for 279 Electoral Votes.
Even: (likelihood 65%) I view this turnout model to be the most likely. Obama has a small lead in early voting that is overwhelmed by GOP turnout on Tuesday. Obama is still able to get his core supporters to the polls, but it isn't enough to compensate for a very fired up GOP base and the Independents. One factor that everyone is missing is the Evangelical vote. As I noted in the IBD/Tipp poll, religious votes are favoring Romney by a large number. This block is able to produce 25 million votes for the GOP, when it votes.
One of the things that everyone misses that the most played commercial of the Obama campaign is footage of Romney saying that he opposes Roe v Wade and will defund Planned Parenthood. This might be a good commercial for the Obama base and the "War on Women", but it also drives evangelicals to the polls and makes them enthused to vote for Romney. This ad may end up being the worst decision of the Obama campaign.
Using this model Romney will win 50.38% to 46.52%
With this 3.85% win, Romney will take FL, VA, CO, NH, OH, WI, IA, and PA for 315 Electoral Votes.
R+4.2: (likelihood 25%) This is the turnout to end all turnouts. The irony is that all it take to believe this model is to believe the 3 month average of a poll of 45,000 respondents with a MOE of 1, is actually correct. The average of the Rasmussen Party ID poll for the last 3 months is R+4.2%. Even Gallup is finding that R+3 is possible, with leaners. In this turnout model the GOP gets all of the self identified Republicans, and the reports that Obama has been cannibalizing his election day voters turns out to be true. Turnout among Democrats is way down from 2008.
Using this model Romney will win 51.96% to 44.94%
With this 7.02% win, Romney will take FL, VA, CO, NH, OH, WI, IA, PA, MI, NV, OR, MN and ME-2 for 355 Electoral Votes.