Wednesday, October 3, 2012

If the Election Were Held Today, Mitt Romney Would Be Elected President of the United States

I know the meme that has been going around is that the Romney campaign is in trouble and Obama has this thing won.  Nothing could be further from the truth.  People are basing this conclusion on polls that are using 2008 as a basis for their turnout estimates, which is just not going to happen. 

2008 is what is called a "wave" year where one party has a tsunami of votes that pushes them to victory across the board.  Besides winning the Presidency, the Democrats were able to pick up 21 additional House seats and a whopping 8 Senate seats, giving them a filibuster proof majority.  This came on the heels of a two term Republican presidency in which the incumbent became very unpopular due to a financial meltdown, which coincided almost exactly with the election, and the ongoing angst over the war in Iraq.  It also didn't help the Republicans that their nominees for President and Vice President hardly exuded competence or readiness for the job.  Besides Sarah Palin's disastrous media interviews there was that whole thing with McCain suspending his campaign to work on the financial crisis, then not doing anything.

For obvious reasons, 2012 is very different.  This time the Democrats are reaping what they have sowed with a 4 year recession, an inconceivably high debt and regulations that are felt nationwide.  Based on Gallup data, there have been some major shifts from 2008 in terms of party identification.  Let's take a look (h/t Michael Franc):

Party identifications shift as does turnout.  You can't assume that an election will look like the last one.  In 2004 in Ohio, Republicans had a 5% advantage over Democrats, which turned into an 8% deficit in 2008, that's a 13 point swing in 4 years!

As someone who works with models all the time (excel not fashion), one thing I've learned is that it is garbage in, garbage out.  If you want to get a certain answer you almost always can get it, irrespective of the underlying data.  In the polls we are seeing, it almost doesn't matter what people say, what is driving the results are the party turnout assumptions made by the analysts.

What you have to do to combat this is look at the underlying data itself and try to make adjustments to the assumptions.  Let's take a look at some polls.  In the National Journal poll that came out today, Romney and Obama were both tied at 47%, however this was based on a Democratic advantage of 7 points over Republicans essentially in line with 2008, which is obviously not going to happen.  Also, Romney is leading Obama by a full 8 points amongst Independents.  It's not hard to see that if Republican and Democratic voting is even close to parity, Romney is going to win this thing, and probably by 5-7%

I realize the National Journal poll is just one poll, so let's look at another one, CNN, which showed that Obama supposedly had a 3 point lead over Romney, 50-47.  Once again, they used 2008 as their model with an 8 point Democrat voter advantage but once again Romney is winning amongst Independents by 8%. 

What about the state polls you ask?  Most of them are highly questionable with methodologies all over the place.  As we have seen from the Gallup table above, party identifications have shifted markedly since 2008 so as most pollsters are using 2008 numbers they are very wrong.  In fact, in some cases pollsters seem to be expecting a year even worse for Republicans than 2008, which is laughable.  In the Columbia Dispatch poll of Ohio voters, only 43% of their sample voted for McCain in 2008.  He took Ohio with 47%.  So that knocks off 4% off of Obama's 9% lead right there!  If you go state by state, you see similar issues.  In a We Ask America poll of Nevada voters, Obama supposedly has a 10 point lead over Romney despite the fact that Romney is actually winning independents by 15 points!  Given that party identification is almost at parity according to Gallup, I'd put Nevada in Romney's column.

And let's not forget the lesson from the Wisconsin recall election.  The Wisconsin exit polls were a great reminder how wrong polls can be.  Remember that they showed Walker tied with his opponent while in actuality he won by 7 whole points.  These polls were conducted the exact same day as the election and interviewed people as they were leaving the voting booth.  This means that you wouldn't get any skew due to people saying that they are likely voters when instead they are playing their PS3 on election day and yet it was still off by a whole 7 points!  

If Mitt Romney wins, I'm sure there will be liberals going nuts shouting how the Republicans stole the election because the results don't match the polls.  Well guess what, that is what happens when you take polls at face value and don't look at the details.

Now, let's just hope he does well tonight so he can make this election a real landslide.


  1. If America only knew....

    Will America be lost because noone had the guts to expose Obama?

    The best Obama exposure site on the net:
    (by our own FReeper, Beckwith)

    The Obama File

    The United States Library of Congress has selected for inclusion in its historic collection
    of Internet materials

    Just a few of the pages: