Monday, October 08, 2012


Look, the new Pew poll caught the race in the middle of a Romney bounce and a wave of Romney enthusiasm, which is why a lot more Republicans were willing to answer Pew's questions than in previous polls, and why Romney leads the poll by 4. Gallup had a poll out today showing Obama up by 5, and, as Kos notes, even Rasmussen shows Romney coming back to earth -- but the press and blogosphere and Twittersphere are ignoring those polls and obsessing over Pew.

This just tells me that the press is, for the moment, on Romney's side -- that's the real story here. Yes, I think Romney got a bounce, but I think the bounce brought the race more or less to parity, which is where it is in the Real Clear Politics poll average. (Talking Points Memo still has Obama up by 2.7 -- and oh, look, Mitt's still down by nearly 4 in RAND's multiple-decimal-point poll.)* I'm sorry, but Romney hasn't done anything to build on that strong debate -- his foreign policy speech was a flop, and no one except his base wants to hear a Republican rabble sabers. It will worry me if we're still talking Obama's-in-trouble by the end of the week, and it will worry me if Obama and Biden have any more bad debates. But Mitt Romney is still a weak candidate. Obama's advertising and messaging have defined him as such, as have Romney's own words and deeds. Polls in the next few days should confirm that.

ALSO, TOO: This.


*UPDATE: I had this wrong -- TPM has Romney up by 2.7. But take out the one Pew poll and it's Obama up by 2.1. I don't understand the weighting -- the Pew poll is the only poll of the past seven listed at TPM that has Romney in the lead, and yet Romney's in the lead. That tells me TPM's average will settle to parity or an Obama lead later this week.


Danp said...

Reinforcing my opinion that polls are more about driving opinion than measuring them. If you are a pollster and want media attention, there is temptation and numerous ways to manipulate them. Only your programmer needs to know for sure.

Jack said...

Shortly after this post went up, TPM updated to reflect Romney up by 2.7.

One debate and now we have those on our side sounding like the Republicans, bitching about polls and their bias.

Victor said...

Follow the money.

The MSM wanted a close race, so they could squeeze every single last penny out of the billions of candidates, PAC, and Super PAC, bucks.

Mission accomplished!

Steve M. said...

One debate and now we have those on our side sounding like the Republicans, bitching about polls and their bias.

I'm talking about putting data points in perspective. I just don't think Romney's really up by 4. I think it became a tied race, or a race with Romney slightly ahead, during the debate afterglow. I think it'll settle down to a 1- or 2-point Obama lead soon, unless there are more bad Democratic debates. And, yeah, I think the enthusiastic party generates more people willing to talk to pollsters. That was true on our side recently, too. We probably never should have believed the really excellent pro-Obama polls.

What I question about Republican poll trutherism is the notion that party breakdown of the electorate in a presidential election year will look more like that in 2010. It won't. It may not look exactly like '08, but it'll look more like that than like a year when Republicans were fired up and Democrats were utterly demoralized.

Danp said...

Jack - I've been arguing that polls are easy to manipulate and use for purposes of driving opinion rather than measuring it for years. But you show me a news outlet that relies a "most trusted" pollster other than their own, or that consistently shows outliers in both directions, and we'll talk. In the meantime, I will argue that only those nearest election day are testable, and for some reason they tend to come together right about then. We've seen the influence of money in accounting firms (Anderson, to name just one), ratings agencies and the media. Why would anyone assume pollsters are pure academics?

Victor said...

They're not academics - or, many of them are not.

I did polling for a Republican firm for about a year, and you should see how the questions are framed.
You don't need "push-polling," when a gentle nudge will do.

Alatea said...

We all knew the media were getting worried that Obama was getting too far ahead and they wouldn't have anything to blather on about any more - so of course they've seized on Romney's debate "victory" and any good polls for him. It gives them something to do and justifies their salaries.

We still have four weeks to go - Romney will screw up again - he's a congenital liar - and a bad one - and he'll forget what he said last week or last month and get stung again. Plus Ryan is like the Hulk - just waiting for the spark that will turn him into a huge green rage monster.

Danp said...

Victor, there is a science to polling, though. And I would argue you don't even have to frame the questions in devious ways. Just program the dialers to target certain demographics not reflected in your internals (e.g. rural vs urban).

Never Ben Better said...

Well, CNN is orgasming today about how it's a miracle, a miracle I tells ya! and Romney's made it a HORSERACE again!

* pant pant pant *

BH said...

CNN et al. can hyperventilate all they want to, but my eye's on Nate Silver's stuff. Until Nate starts talking toss-up - and in the Electoral College count, not a meaningless national head-to-head - I count the rest as just showbiz.

Jack said...

Steve M,
I should have been more clear, but I wasn't referring to you when I mentioned "people on our side." I was just thinking about the various sorts I encountered around the internet last night who are in denial about what's happening.

All I know is that the mid-term future (next 30-50 years) looks very different depending on who wins this critical election. Any sane person should be terrified at the prospect that a completely insane political movement dominated by billionaires and American Taliban are about to seize control of the White House, and solidify their control over Congress. There are literally millions of lives at stake: Given the monumental changes to Social Security, Medicare, and Medicaid proposed by Ryan, et al., there are millions of people who will lose their lives years or even decades earlier than they would otherwise.

This assumes access to health care determines life span -- a very safe assumption.

Steve M. said...

Sorry if I misunderstood you, Jack. And I agree with everything you just wrote.

: smintheus :: said...

Dig into the Pew polls internals some more. It's grossly unrepresentative of the population. It massively oversampled older whites, massively undersampled Hispanics. That explains the massive (15%) shift in party identification since the last Pew poll (which fwiw oversampled Democrats).

All the polls that publish internals show the same result: both Romney's and Obama's share of the vote among Republicans, Independents, and Democrats has remained pretty steady despite the debate. So any swings in the polls are artifacts of changing demographics of those who are answering the pollsters' questions.