Sunday, November 12, 2006

Exit Poll Numbers and Thoughts on Turnout

First off, Pho’s Akron Pages is declaring Lori the winner of the Voter Turnout Office Pool. She will be offered a post shortly. Be forewarned; she may write something nice about the Steelers just to annoy me.

Second, if you didn’t click through the link before, YellowDogSammy was the first to point out that the SoS site revised the turnout numbers for Cuyahoga County upward. Cuy. Co. turned out in the mid-50’s – about average for the state.

Also, Bill Callahan commented to a previous post with these trenchant insights on urban turnout.

But mostly this post is about playing with – playing with, not analyzing – exit poll numbers. The poll links come courtesy of, a site set up by MacWilliams Robinson & Partners, Ted Strickland's media production consultants. They plan on maintaining the Battle for Ohio site will at least until 2008.

The exit polls themselves are from CNN, and 2004 is still up, giving us the numbers to play with. In terms of percentage of the vote, Dems and Republicans nearly reversed themselves. In 2004, Dems represented 35% of the vote for President; Republicans 40%. In 2006, Dems were 40% of the vote for Governor; Republicans 37%.

From there we get into deeper, but increasingly shaky analysis. I used the percentages from the exit polls and the vote totals from SoS to calculate the approximate number of Republicans vs. Democrats voting in 2004 Presidential and 2006 Gubernatorial races. Then I took the presidential vote as a baseline for motivated voters in each party, and figured out the percentage turnout for each party. That is, what percentage of the Kerry vote does the Dem vote in 2006 represent and what percentage of the Bush vote does the Repub vote in '06 represent. Here's what we find:


Total votes: 5,722,443

Dem Votes: (35%) 2,002,855

Repub Votes: (40%) 2,266,977


Total votes: 4,177,498

Dem votes : (40%) 1,670,999

    % of Kerry vote turned out: 83.4
Repub Votes: (37%) 1,545,674
    % of Bush Vote turned out: 68.3
Don't bother telling me what's wrong with the methodology. Of course not everyone who voted Dem in '06 bothered voting in '04 or was even in the state or was a Dem or whatever. Not to mention problems with exit polls generally, particularly in 2004. I know.

These are very rough calculations. But they crudely indicate a wide disparity, suggesting that Democrats turned out a much higher percentage of their motivated voters than Republicans did. And this jibes with anecdotal evidence we all heard before and after the elections.

Finally I would note that the 2006 exit poll results pretty much jibe with the final result. It's hard to tell for sure since the CNN page is all breakdown with an aggregate for each candidate. But looking for instance at the results for men and women, it certainly looks like the average is something close to the 37/60 split in the final result. I'm sure Stolen Election Guy will come up with some explanation, but it may take a while.


Gloria Ferris said...


It also doesn't take into account the crossover Republican voters this election, and if what people said was true and I don't have any evidence to doubt it many people crossovered this election.

Personally, I think that we had crossover from Dem to Rep in Cuyahoga County given the fact the Joan Synnenber won against Christine Russo ( a strong Dem name) and so did Dick Ambrose against a Gallagher another stron dem name up north here.

To me, this means and educated voter base that was not fooled by hype and refused to buy into name games.

Good for the public, not so good for the old boys clubs.