The mathematics of Romney's victory

I've been predicting a Romney victory for quite some time now.  This has been based on a stated distrust of the polls: specifically, an overweighting of Democrats in the samples.  I've been using a (made up) rule of thumb that the polls overstate Obama's support by 5-7%.

Well, here's proof, in hard numbers:

Of all the polls you have seen, there is one poll that has gotten no attention, it is a poll that has been taken monthly, it is a poll that Doug Ross spotted and promoted on his site. It is the Rasmussen Poll of party identification.

They have party identification results online dating back to 2004. Here are the 2012 numbers through August:

Interestingly (and in great contrast with almost all of the polls you see), this post shows the same poll for 2004, 2006, 2008, and 2010 so there is historical perspective.  It compares the results of the elections in those years to the party affiliation polls from those years.

The short answer: the election results matched the poll predictions precisely.  You really need to RTWT.

Now look at the average of this year's polls, which look like they sample D+5 on average.  That's exactly backwards - according to this, they should be sampling R+9.  So rather than Obama up by 3 (margin of error), Romney is actually up by 6 (way outside the MoE).  And quite frankly, my prediction of Romney 346, Obama 192 is looking like I'm considerably overstating Obama's chances.

There are of course two issues here:

1. This is a national poll, while the returns that matter are State by State (Electoral College).  We can extrapolate this to the States, but we're kind of going a bit out on a limb.  Of course, that's exactly what I did in my prediction of 346-192 Romney.

2. The votes that matter are the ones that are cast in November, and party identification is quite frankly not as important as turnout.  The post I linked to does a very good job of covering historical turnout compared to these polls, but once again we're extrapolating - as will happen when you're talking about any poll.

At the end of the day, I'm confident in these numbers.  There's plenty of corroboration - polls of whether the country is on the right track or the wrong track reliably come in with large majorities saying "wrong  track" (the latest is something like 55 wrong track vs. 39 right track).  Perhaps two thirds of respondents say that the state of the economy is the primary thing they will consider at the polls this year.  None of this is good for Obama.

Add in my take that the Media has pretty well destroyed their credibility - trust in the media has been declining for more than a third of a century - the only rational conclusion is that this will be a landslide year for Republicans.  And so, I'm updating my projection to Romney 359, Obama 179.  This is based on a 10% shift of current polling numbers at Election Projection.  That means that only "Strong Obama" and "Solid Obama" actually go (D), and everything else goes (R).

Or more properly said, 179 Electoral Votes are won by Obama, and 359 E.V.s are won by "Not Obama", because this election is - like 2010 - a referendum on the Administration of Barack Obama.  I quite frankly don't believe that 37% of voters identify as Republican in any meaningful way.  I absolutely can believe that they identify as someone energized to vote the current occupant of 1600 Pennsylvania Ave out.

So there you have it - predictions, now with moar data.  I may be wrong here, but I've tried to show my work.  You can make your own electoral map, if you want.

Note that I do not say that a Romney victory (especially of this magnitude) would be A Good Thing.  On the contrary.  But the numbers are the numbers, and they don't look good for the Democrats.