A poll conundrum

By Libby

I've never given much weight to polling in trying to predict long term trends. Polls by their nature take a snapshot of a moment in time and by the time they're published some new factor has likely as not already swayed public opinion in another direction. But still, the polling in PA has puzzled me because it seems so far removed from the reality on the ground. I mean consider this recent rally for Obama in Philly. By all accounts at least 35,000 people showed up and their enthusiasm didn't end with the rally.

5,000 people (at least) had nowhere to go but up Market Street. Obama's charge of the night: "Declare independence!" was with them. They started with the familiar "O-Bam-A." By 7th and Market, they had graduated to "Yes we can!" By 10th and Market, with hundreds streaming in between cars on the road, they were just cheering. At first, a few Philly cops, killjoys, tried to rough the crowd to the sidewalks. It didn't work. The cops retreated to the sidewalks. By the time I ducked into my hotel, a full mile away from Independence Park, the Obama crowd was still marching. [...]

I counted at least a hundred Philadelphia police officers. There were state troops. TSA personnel magging the crowd. A helicopter hovered over the square. The fire department set up a command post with extra medical supplies. It was some way to start Obama's final Pennsylvania push.

That sounds more like the aftermath of the World Series win in Boston than a simple political event. Meanwhile, Clinton is drawing mere hundreds to her rallies, yet the pollsters show her leading in PA.

Hillary Clinton leads among bowlers, gun owners and hunters in Pennsylvania, a blue-collar trifecta that is helping her hold an edge over rival Barack Obama heading into Tuesday's pivotal primary there.

Overall, Clinton leads Obama by a margin of 48-43 percent, with 8 percent still undecided. The telelphone survey of 625 likely Pennsylvania voters was taken April 17-18 and had an error margin of plus or minus 4 percentage points.

Do Obama supporters never answer the phone? Are bowlers, gun owners and hunters too busy to go to Clinton's rallies? Clearly, judging by the fundraising figures, they don't contribute money to the campaign either. So what explains the discrepancy between the close predictions and the obvious disparity between the public show of support?

Update: On another note, via Balloon Juice commenter Martin, an uncommited superdelegate is asking people to vote on who he should support. The tally when I checked was 98% in favor of Obama. Granted, it doesn't mean much in terms of the PA vote since anyone can vote in it, but still it indicates a level of support that I see consistently in news reports about these rallies that never seems to be reflected in the polls.

Comments

[Stacie Boschma]

I think some of it is the urban/rural split. Obama does especially well in urban areas, which can churn out crowds of thousands, while less urbanized areas just have fewer people to turn out for rallies.

I think one thing that surprises me is Clinton's relative strength among more rural voters. She's an establishment millionaire decades removed from the concerns of ordinary folks. Not saying Obama's Chicago years and Harvard education should play better, I've just been surprised by Clinton's hold on those areas.

[Libby]

Yeah, I thought of that Stacie but it's not just happening in Philly or even just in PA. Obama consistently outdraws her everywhere that I've seen, even in small towns.

I just don't think the polls reflect the public sentiment generally.

[BJ]

Polling results are based on statistical models based on turnout assumptions for certain voting blocks. If some of those voting blocks wind up turning out in much larger, (or smaller), numbers than they've been modeled to, then the polls won't match the reality. Unfortunately, it is nearly impossible to predict ahead of time that kind of a shift in demographic voting behaviour.

Anyway, Booman's got an interesting take on the turnout question.
http://www.boomantribune.com/story/2008/4/20/124859/915

You'll have to ask your resident Pennsylvanian whether or not that makes sense.

[Libby]

Thanks for the link BJ. That is an interesting analysis and one I'd tend to agree with even not knowing the state very well but it's not unlike the demo in MA. The urban areas are concentrated in small areas but they're voter heavy and will trump even a wide spread win in the rest of the state. The major power is also split between Boston which is white collar heavy and a couple of the larger urban centers in the East that are blue collar. In MA the urban votes are controlled by the local machine or the unions. I have no idea how it works in PA, but I bet Fester is going to tell us eventually.

[Libby]

Shoot, I hit post too soon. My main point is that the polls are not reflecting the numbers we can quantify, like how many people show up to rallies. I don't care if it is a big city, 35,000 people is a lot of support. When is the last time you ever heard of that many people showing up and then marching away from a proforma event like that, freaking chanting?

You can't get that many people to a antiwar rally anymore and 2/3 of the country opposes the war. I've been wondering about these close turnouts in the big key states when almost all the smaller ones are blowouts for Obama.

I have this uneasy feeling about it. The states Hillary has barely won are the same states that the GOP manipulated to steal the 04 election. The polls were consistently wrong then too. It's good to remember the GOP still controls the voting tally and they want Hillary. I'm sure they anticipated her getting the nod and have their oppo locked and ready to fire and we didn't solve the verified voting problem by a long shot. And you've noticed that Rove has finally crawled out from under his rock and he's just as cheery as pixie in springtime what with the Dems doing his work for him.

It feels like we're being played to me. It wouldn't take much work to come up with a really good conspiracy theory about it.

[BJ]

I guess I'm not as conspiracy minded as you are, Libby. The early polls were off due to bad turn-out projections, which the pollsters adjusted and got better results in later states. Even then, the only real shock seemed to be New Hampshire.

Obama came from a huge national deficit to basically tie Clinton overall on Super Tuesday, which doesn't sound too suspicious if you think the GOP is trying to help Clinton. Post-Super Tuesday, the only states she's won has been Ohio, Rhode Island, and Texas, kind of, all states she was expected to win, and all states where Obama was able to tighten the margins.

If Hillary beats the polls in this one and scores a blowout, then I'd be wondering. Until then, I'll content myself with the right-wing conpiracy not being as vast in the Democratic primaries.

[Annole]

I don't know if this explains it, but it gave me a hmmmm moment when I read it.

From McClatchy

HOW WE POLL

“…..Those interviewed were selected by a random variation of telephone numbers from a cross section of telephone exchanges. That means anyone in the state with a phone line had the same odds of being called as anyone else, except for people who use cell phones only. Cell phone numbers are not in the exchanges….”

I found that part about cell phones particularly interesting. I know many people, especially young people, who not long have "landlines".

[Annole]

Sorry BJ. Somehow my post (see above) listed under your name. I have no idea how.

[Libby]

LOL BJ. I'm not putting out any theories yet, but I don't find it out of the realm of possibility either. I never underestimate the long range thinking of Rove and I believe he's still very much involved in strategy for the GOP.

Annole, that's a good point and one that I'm actually aware of which I think is a big flaw in the polling system presently. I don't think they reach new voters well and those are the ones that are driving Obama's success.

By the way, don't be confused by the labeling system on the comments. Your name appears under the comment, not above, so it is correctly attributed.

[TW]

It still amazes me that Obama, who turned his back on who knows how many incredibly lucrative offers from elite law firms to work as a community organizer on the South Side of Chicago, can be pigeon-holed as elitist and disdainful of working class people. I recognize he's to blame for not pushing back on this hard enough or effectively enough, but the audacity of the allegation really sticks in my craw, especially coming from the likes of Sen. Clinton, Chris Matthews, or any of the conservative talking-heads.

What really saddens me about it is that i think it really does come down to race. Just objectively, the only thing Hillary has more in common with white working class people than Obama is the color of her skin. When Clinton partisans speak of Obama's unelectability, they can't point to much in the way of evidence. (So he's not equally strong in every demographic - so what, no politician ever has been; and if his vulnerabilities were crippling, how has he managed to win more votes so far than anyone in either party, including McCain?) What they mean - and what their surge in the last few states has been driven by - is that when faced with the strong likelihood that a black person will win the nomination or the presidency, a lot of white people will get very hesitant; and that it won't take much innuendo or suggestion to push them to vote for the alternative, whoever that alternative is.

So there's a challenge ahead for Obama. But it's really a challenge for all of us, too.

[Libby]

Well said TW. I also believe that racism is roiling just under the surface. I'm one of the few people that believed Hillary took New Hampshire because a lot of cranky Yankees just couldn't bring themselves to vote for a black man. I'm not sure it's even a conscious thing. Blue collar types are likely as not to have black friends, and on that basis will deny vehemently that they're racist but then in the next breath will broadbrush all young blacks as thugs wearing pants down to their knees.

It exists in white collar liberals as well of course, but it's much less overt and many perhaps even aren't aware of certain ingrained biases that rest deep in the psyche and don't realize they're susceptible to subliminal messaging.