Wednesday, October 10, 2012

October 10, 2012 Ohio Analysis

Did you know I have a model for Ohio as well?  You probably didn't.  I use the same methodology I use for the national numbers, then adjust for different historical turnout models.  Let me talk for a moment about the models.

Ohio has swung a great deal over the last few years.  In 2004 they were R+5, but in 2008 they were D+8.  I am using 4 models in this reweight, 2004, 2008, 2010, and 2012.  The 2012 number is based on registration estimates made by a couple professors at Ohio State.  Ohio does not have party identification, so you can't look the current party split up on the Secretary of State web site.  So take the 2012 model with a grain of salt.

2004 D/R/I - 35/40/25 - R+5
2008 D/R/I - 39/31/30 - D+8
2010 D/R/I - 36/37/27 - R+1
2012 D/R/I - 37/37/26 - Even

Polls included:
O+1 - Survey USA
R+1 - ARG
R+1 - WeAsk
O+1 - Rasmussen

O+0.80% - Current RCP Average
R+7.77% - Average using the 2004 turnout model
O+1.68% - Average using the 2008 turnout model
R+4.56% - Average using the 2010 turnout model
R+3.84% - Average using the 2012 registration model

If Ohio turnout is similar to 2010, then Ohio will easily turn red this year.  Even if turnout is similar to 2008, then Obama would barely hold the state.


  1. Dave,

    Great analysis.

    Ohio is what keeps me up at night. I tend to think Romney will pull it out, but if the same people who show up in 2008 show up in 2012, it looks like Romney will narrowly lose. One interesting fact is, Obama didn't win Ohio by a large margin in 2008, if something like 2.5% of voters changed their mind, McCain would have won.

    What is also comforting is around 500k voters were dropped from the voter rolls and I believe they now also have a Voter ID law. I see that shaving quite a bit off of Obama's totals.

    At the end of the day, Ohio (like Virginia) is traditionally a Republican-leaning state, and both swung solidly back to Republicans post-Obama in off year elections. In fact, they were blowouts, so all this talk about them being "new" blue states was nonsense.

    If you're looking for an article to write, I'd love to see some likely scenarios of how Romney gets to 270 without Ohio. My guess is he'd have to win Colorado, Iowa, and New Hampshire.

  2. I keep an eye on the Ohio absentee ballot request spreadsheet and this has been telling us for weeks that the voting pattern is not like 2008. R's are behind in Cuyahoga county by 2:1 compared to 3:1 in 2008. Clark county, described as the swingiest county, went 2:1 for D's, now about 2:1 for R's. Warren county, a Cincy suburb, an R suburb is going 7:1 R, showing a big enthusiasm factor.

    Cuyahoga county is the key. If Obama doesn't come out of that county like he did 3:1 in 2008, he's going to get overpowered in all the other counties showing much increased R ballot requests.

    Another way to look at this came from a polling company, who looked at the large voter registration drop in Cuyahoga and concluded a Romney win in Ohio:

    I guess we'll have to see, but Ohio has been looking much more promising for Romney than the pre-debate polls were saying.

    1. Its evidence the poll sample models may be wrong, but I wonder if there's more going on.

      Don't get me wrong, it's a nice trend, but I'd like to know more about other variables that might have changed.

  3. There's more going on......

    Apparently new in 2010 was all registered voters received an absentee ballot request form in the mail. Previously it was the big democratic counties that sent these out in 2008. So now smaller presumably 'R' counties are getting absentee request forms.

    I've got an idea for a new analysis, I'm going to run some numbers through excel tonight and report back.

  4. Here's a google spreadsheet of analyis done to see if the Ohio electorate will be similar to 2010. the answer is a little bit more 'R'

    1. Wow, that is some excellent work! I'm going to steal some info from in this for my next Ohio update.

      Absentee ballot requests are down by 7.9% from 2010 in Cuyahoga County? That is devastating for Obama. He has to run up the numbers there to overcome the Republican increases from downstate.

  5. They are not really down 7.9%, they are down 7.9% relative to the rest of the state. More of the effect is likely, other parts of the state are up. But that's why I think I've found something in this comparison.

    Now it could be that absentee voting is a trend established by urban counties and is spreading outward. That's probably some of it as well, but the way this looks is that the electorate in Ohio is going to be a repeat of 2010 at least.

  6. Dave, have you thought about opening up the comments for instant publish rather than moderated?