Friday, September 28, 2012

Romney's 17 point lead w/ Indys in PPPs OHIO poll could be his saving grace

The NRDC Action Fund must have been thrilled after receiving the results from PPP's swing-state project. Obama leads. . . well, everywhere, and by strong margins. Even in states that have been historically very close, he seems to be pulling away. That is the case with PPP's newly released Ohio poll (note the poll was taken September 14-18, 2012). The President leads Mitt Romney 50-44%, the widest margin between the two candidates since May of 2012, when Mitt trailed Obama in Ohio 50-43%. But as always, the devil lies in the details, or, erhh, the crosstabs.

How exactly is Obama leading by 6 points in a state John McCain only managed to lose by 4.6 points? Because Obama is attracting a larger portion of his Democratic base (89%), than Romney (85%), and because 8% MORE Democrats than Republicans now say they'll show up to vote in Ohio in November. To be precise, PPP finds Ohio voters self identifying as 41% Democrat, 33% Republican, and 26% Independent, or D+8. That is the SAME net margin of Democrats to Republicans that voted in Ohio four years ago, when party ID was 39D/31R/30I. But much like the Virginia and Pennsylvania polls discussed here and here, Romney is destroying Obama amongst Independent voters. In Ohio, he carries them by 17 points, 51-34%. So although the PPP partisan ID findings are practically the same as 2008 exit poll results, lets re-weight the poll to fit 2008 turnout. The main reason i'm choosing to do this is because PPP finds Independent voters making up 4% less of the electorate this November than they did in 2008 (26% vs. 30%). Because Romney's lead among Independents is so large, I'm curious to see how the '08 numbers would change the poll result, if at all:
PPP poll weighted to 2008 CNN Ohio exit polling:

Party ID        Obama                 Romney

39D               34.71 (89%)        4.29    (11%)
31R                2.48   (8%)          26.35 (85%)
30I                  10.2  (34%)        15.3    (51%)
                       47.39%                45.94%

As you can see, that four point drop in Independents that PPP found in the Ohio electorate is what makes the difference. If you take PPPs crosstabs and apply those numbers to a 2008 turnout scenario, the6 point, 50-44% Obama lead becomes an insignificant 47-46% lead, all on the strength of Romney's support among Independents.
And unsurprisingly, Mitt Romney takes the lead, even using PPP's numbers, when re-weighted to reflect a 2010 style electorate (36D/36R/28I):

PPP poll weighted to 2010 CBS Ohio exit polling:

Party ID        Obama                 Romney

36D               32.04 (89%)         3.96    (11%)
36R                2.88   (8%)          30.6    (85%)
28I                 9.52  (34%)         14.28   (51%)
                       44.44%               48.84%

In other words, Romney is sitting pretty in Ohio if the electorate looks similar to two years ago. He would lead 49-44%. But if the Ohio electorate looks more like 2008, as most polls are finding, Obama's lead is still only 47-46% because of Romney's surprising strength with Independents. Since the electorate, in my opinion, isn't very likely to look like 2008 or 2010, lets average the turnouts from those years, and re-weight the PPP Ohio poll to that partyID (37.5D/33.5R/29I):
PPP poll weighted to average of 2010 and 2008 Ohio exit polling:

Party ID        Obama                 Romney

37.5D            33.375 (89%)       4.125   (11%)
33.5R             2.68   (8%)          28.475  (85%)
29I                 9.86  (34%)         14.79   (51%)
                       45.915%               47.39%


When you consider the PPP numbers in light of a reasonably fair turnout scenario this November (D+4, the average between '08 and '10 turnout), then Romney actually leads in Ohio, 47-46% (something we haven't seen in a long while.) In other words, if Romney is winning Independents on election day by 17 points, as he is in this poll, then it is more likely than not he will carry the state.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.