10/12/2012

Ohio Poll Magic And Reality

From-Jer's Notes

I thought I would take a bit of time to explain why I'm not really worried about the election and the best way to do this is to focus on the state which will be the key to the election-Ohio.

Despite all the talk Romney could win the election without winning Ohio, in a multiple of ways actually. But the fact remains that if he doesn't win Ohio the chances are that he is not going to win the other states that he needs to win without Ohio. Though I will say that if it is razor thin in Ohio that might not necessarily be the case, for example Romney may win Wisconsin and Iowa with razor thin margins and offset a razor thin loss in Ohio. Regardless Ohio is important.

So let's take a minute to study Ohio's polls since the debate. Unfortunately I do not have Rasmussen's internals where Obama leads by 1% so I'll leave them out.

Let's start with the  NBC/Marist poll which shows Obama up by 6%

Notice that since the debate Romney has picked up just 2% while Obama has remained at 51%. But the question is, are we comparing apples to apples in the two polls? Actually not at all. here are the party ID numbers from the above polls from NBC.


Notice that in the October 11 poll where Obama stayed the same and Romney gained 2% for an Obama +6 advantage, the party affiliation was D+11 (40/29/29) whereas the poll just a week before where Obama was supposedly leading by 8% had a party affiliation of D+5 (36/31/32). So on Oct 11, NBC/Marist had to poll 4% more Democrats and 2% less Republicans in order for Obama to lose only 2% in their poll. 

Not only did Romney pick up 2% on Obama in a week according to this poll he did it under longer odds (so to speak) than just a week before. It took an extraordinary shift in party affiliation by the pollster to achieve the result they wanted.

Now let's compare two other polls done basically at the same time in Ohio. In fact this first one interviewed more than the NBC poll and thus has a smaller margin of error.

It is from Gravis which actually shows Romney up by almost 1% among likely voters in Ohio.


So we have a poll (NBC/Marist) showing Obama up by 6% with a D+11 turnout model and you have Gravis showing Romney up by 1%  with a D+6 turnout model.





Amazing what happens when you take a 5% turnout advantage away from the polling. That extra  5% advantage in the NBC/Marist polls takes a race from a decisive victory for Obama in Ohio, where by the way he only won in 2008 by 4% with a D+8 actual turnout advantage and makes it a dead even race. The difference between a D+11 and a D+6 turnout in Ohio very probably could be the difference between a President Obama and a President Romney. Think NBC knows that?

To show how close this could be let's look at one more recent poll in Ohio from Survey USA, which gives Obama a 1% lead.


This poll too has a D+6 turnout model (Bottom line of chart)


So based on these polls we can make the following observation, the break even model turnout for Ohio is probably somewhere around a D+6, In other words if 6% more Democrats than Republicans vote in Ohio, then the race could go either way. If more Democrats vote as in the absurd NBC/Marist scenario then Obama will win and if Democrats have less than a 6% advantage, Romney will win.
This is backed up by the actual vote in 2008 when Obama had a D+8 and won by 4%.

So the question becomes will Ohio have a D+6 or above turnout or less? Well we do have  indications that the early voting which favors the Democrats in Ohio is not nearly so lopsided in 2012 as it was in 2008. Add to that the party enthusiasm gap between the parties has basically reversed since 2008. But the real question is. is it possible for Romney to seriously cut into that magical D+6 in Ohio? After all in 2008 Obama had that D+8  can Romney and the Republican possibly shave tree points off Obama's turnout lead?

Well as it turns out it is not only possible it is very likely,  since you see in 2004 Bush won Ohio with a R+5 turnout. That's right in one election cycle the turnout in Ohio shifted 13 points from Republican to Democrat and we are to believe that with all that has happened, with all the angst that Republicans and Independents have towards Obama that he is not going to lose three percent of his turnout base in Ohio?

In 2008 Obama won Independents by 8% . Note the last chart above, Romney is winning Independents by 9% in Ohio  a net shift of 17% against Obama. Reality is that Obama is losing the Independent vote, Republicans are eager to vote him out, and Democrats are not nearly as enthused to vote for him as they were four years ago. All of this adds up to a Romney victory in Ohio and the Presidency.

I won't say take it to the bank, but I'm seriously considering taking it to Intrade and making a few bucks. 

2 comments:

  1. the biggest elephant in the room...

    I'm 2008 only 30% of the republicans voted..
    ...and 10% voted for Obama

    That is how they get "likely voters"...based on the 2008 turnout

    This time every republican is going to vote, even if they have to crawl...
    ..and there is no white guilt

    ReplyDelete
  2. I've been following your posts about the polls with interest, and I'm reading more articles that support what you've said all along. It's encouraging!

    ReplyDelete