Posted on | October 25, 2012 | 33 Comments
Excuse me for my continued attention to Nate’s graveyard-whistling, but no matter how clear the evidence of a pro-Romney trend, the Grand Swami at the New York Times won’t stop. He’s now raised the likelihood of Obama’s re-election to 71.0%. (The one-tenth of a percentage point being necessary to the pretense of scientific exactitude.)
Is Nate Silver hustling an insider-trading scam with InTrade? Or is he merely acting as an Obama pompom girl? Either way, the poll-watcher at the nation’s most influential newspaper cannot be unaware of how his coverage functions to shape elite opinion, which is in turn reflected in other media coverage that then influences mass opinion, and believing that Nate Silver is acting as an honest neutral broker in this transaction requires a faith in human goodness that I lack.
When I looked at the Real Clear Politics national average this morning, my eye was automatically drawn to the ABC/Washington Post poll that shows Romney now leading 49-48. This strikes me as significant because two weeks ago (Oct. 10-13) the same poll had Obama leading 49-46. Thus they are reporting a net pro-Romney shift of +4. It doesn’t matter for this purpose, by the way, whether ABC/WaPo is accurate as to the actual number, so long as their method is consistent — the trend is what it is. Meanwhile, both Gallup and Rasmussen tracking polls have Romney leading by 3 points, and the latest Associated Press poll has Romney ahead by 2 points.
“But, but, but . . . Ohio!”
Yeah, I know: Obama’s poll numbers in the Key Battleground State of Ohio — as it is always called — are consistently better than in the national polls, but I do not believe that any amount of advertising, campaigning or ground-game organizing in Ohio is going to enable Obama to overcome at a purely local level what is a clear national trend toward Romney.
That is to say, if Mitt’s pulling ahead nationally, Ohio will inexorably move in the same direction, even if the shift is smaller. So when we look at the Ohio polls and see three that show it at tie, and five others showing Obama ahead by 1%-5% margins — a net advantage of 2.1% in the RCP — we have to weigh two facts (a) Obama’s small but persistent lead, and (b) the national trend toward Romney. If the trend continues, Ohio will follow that trend, and subsequent polling will reflect it.
By the way, that Time poll of Ohio is so hinky as to be dubious.
Time hasn’t previously polled Ohio at all this cycle, and for them to come in during the last two weeks of the campaign with a poll on the high side of the margin — matching the Obama +5 of PPP and Quinnipiac — is highly dubious. You can go chew on Time’s sample if you want, but the statistical details are irrelevant to my basic suspicion. If it walks like a scammy poll and talks like a scammy poll, it’s probably not a duck.
We do not know, by the way, whether the Romney trend will continue over the next 11 days. Polls are a lagging indicator and you can’t extrapolate numbers forward. Welcome to the Twilight Zone phase of the campaign, where things become mysterious. But be encouraged by Ed Morrissey’s analysis of the new AP poll:
Now Romney leads without leaners by three points, 45/42 — and a firm number of 42% for a incumbent is a disaster 12 days out from the election. When the “could change mind” voters are taken out, it’s even worse — 43/41 for Romney, with the incumbent barely above four out of ten voters.
Too early for predictions, but an incumbent persistently below 50% in October is usually what polling experts call “a loser.”
UPDATE: Ted Frank offers an informed statistical critique of Nate Silver’s methodology, which is different than my own criticism, namely that sabremetrics doesn’t work for polls, because politics ain’t baseball. Political opinion and voter behavior differ from the performances of baseball players, and “weighting” polls based on past performance cannot overcome the fact that polls are a lagging indicator. Nate Silver’s pretense of scientific exactitude — his constant and increasingly obnoxious reference to his own self-proclaimed expertise — is itself a factor that may shape public opinion, and he ought to damned well know it.