Wednesday, October 31, 2012

A Brutal Polling Day? Not for Mitt Romney

I guess folks on the left don't really look at survey internals, which is odd, since all these drive-by commenters keep saying conservatives can't do math. Looks the other way around, frankly. A good example is this report at TPM, "After Brutal Polling Day, Romney Team Reassures That They’ll Win" (at Memeorandum).

And the evidence for these so-called "brutal" numbers for Romney? Well, an obvious outlier at National Journal that has Obama up 50-45 when no other national poll of any repute shows a spread in Obama's favor like that. Even the hopelessly Democrat-heavy New York Times/Quinnipiac poll today had a miniscule Obama edge at 48-47. And on Monday Pew Research had the race deadlocked at 47 percent, with a turnout edge for Romney.

TPM's also claiming Obama's up by five in Ohio, which is again relying on the Times/Quinnipiac survey for the Buckeye State. But as I noted this morning, NYT's Ohio numbers are badly off, way out of line with both 2008 exit polling data on Democrat and Republican turnout, as well as likely turnout numbers for the GOP this year with the enthusiasm gap taken into consideration. Bryan Preston has more on that, "That Q Poll Showing Obama Up by Five in Ohio Has a Flaw (Updated: PPP Too?)." (PPP is the far-left Kos pollster, which almost always favors Democrats and is thus widely discredited.) See Ed Morrissey on those NYT numbers as well, "Final CBS/NYT/Q-polls in OH, FL, VA show Obama up …":
In each of these three states, the CBS/NYT/Q-poll shows Republicans at a lower percentage level of turnout than in the 2008 election. If one makes that assumption, it’s not too difficult to be guess that Obama might be ahead. However, that’s exactly the opposite of what all other polls rating enthusiasm are telling us what the electorate will look like on Tuesday. In fact, it’s not even what this poll shows, with Republican enthusiasm +16 over Democrats in Florida, +14 in Ohio, and +7 in Virginia.
And to round things off, here's the poll from the University of Cincinnati's Institute for Policy Research on Ohio, which has Obama up 48-46 with a partisan breakdown of D-45, R-43 percent, and I-12 (and note that independents here, who are breaking for Romney in all other polls, are probably under-sampled, to say nothing of the over-sampled Democrats).

So, it's not Mitt Romney who's having a "brutal polling day." If anything, it's the truth that's having a "brutal honesty day." Polling methodology (i.e., math) is not hard. If progressives get it they're not letting on, which is even worse from an integrity standpoint.