Slog News & Arts

Line Out

Music & Nightlife

« A Talk With One of Those Femal... | Las Vegas is the Most Depressi... »

Thursday, January 10, 2008

Why the Polls in New Hampshire Were Wrong

posted by on January 10 at 8:57 AM

I hear there was some grumpiness about my post yesterday wondering whether issues of race warped the pre-primary polls in New Hampshire. Today the president of the Pew Research Center writes in The New York Times that he suspects the polls were warped by… issues of race and class. He discusses the other theories floating around and then writes:

To my mind all these factors deserve further study. But another possible explanation cannot be ignored — the longstanding pattern of pre-election polls overstating support for black candidates among white voters, particularly white voters who are poor.

In exploring this factor, it is useful to look closely at the nature of the constituencies for the two candidates in New Hampshire, which were divided along socio-economic lines.

Mrs. Clinton beat Mr. Obama by 12 points (47 percent to 35 percent) among those with family incomes below $50,000. By contrast, Mr. Obama beat Mrs. Clinton by five points (40 percent to 35 percent) among those earning more than $50,000.

There was an education gap, too. College graduates voted for Mr. Obama 39 percent to 34 percent; Mrs. Clinton won among those who had never attended college, 43 percent to 35 percent.

Of course these are not the only patterns in Mrs. Clinton’s support in New Hampshire. Women rallied to her (something they did not do in Iowa), while men leaned to Mr. Obama. Mrs. Clinton also got stronger support from older voters, while Mr. Obama pulled in more support among younger voters. But gender and age patterns tend not to be as confounding to pollsters as race, which to my mind was a key reason the polls got New Hampshire so wrong.

Poorer, less well-educated white people refuse surveys more often than affluent, better-educated whites. Polls generally adjust their samples for this tendency. But here’s the problem: these whites who do not respond to surveys tend to have more unfavorable views of blacks than respondents who do the interviews.

RSS icon Comments

1

I think the men thing is what makes Obama the most viable in November. Dems have a structural lead with women, so once the nominating process is over, he'll be able to take advantage of that, but if he can also get men to vote for him, particularly white men, the GOP advantage among men won't be as strong and will lead to many a narrow victory.

Posted by Gitai | January 10, 2008 9:14 AM
2

Is it racist to not vote for Obama in the primary because you think "America won't elect a Black president?" even if you think he's the best candidate?

Is it sexist to not vote for Clinton if you think "America won't elect a woman?"

Posted by angel | January 10, 2008 9:17 AM
3

When will we move beyond asking, "why were the polls wrong?" I get the interest, but not sure there is a point to it all, except to remind us, as Tom Brokaw said, that maybe media and the rest of us need to pay less attention to polls, place less value on them, because they do have the ability to sway attitudes, incorrectly perhaps.

Enough already.

Posted by Brad | January 10, 2008 9:27 AM
4

I haven't read the PEW article yet...but they should be able to tell from the crosstabs of their previous polls and the exit polling whether or not his theory holds any water. He just has to compare the group that he claims changed their vote...

I, for some reason, am resistant to the theory of racism. Do we really think this played out better in Iowa than in New Hampshire? That makes little sense to me.

Posted by Timothy | January 10, 2008 9:27 AM
5
But here’s the problem: these whites who do not respond to surveys tend to have more unfavorable views of blacks than respondents who do the interviews.

Erm ... how do they know this, exactly? Did they poll them?

Posted by tsm | January 10, 2008 9:28 AM
6

Eli, your argument yesterday about why race was a factor in the polls error didn't ring true to me, but this one does. It seemed like yesterday you were saying that people will tell pollsters that they will vote for a black candidate and then behind closed doors, vote for the white candidate. This argument is more that the polls themselves don't reach the people who would be prejudiced against a black candidate and don't have an effective way to account for the differences in attitudes.

Posted by Julie | January 10, 2008 9:28 AM
7

Poorer, less well-educated white people refuse surveys more often than affluent, better-educated whites.

these whites who do not respond to surveys tend to have more unfavorable views of blacks than respondents who do the interviews.

And where's the info on how all those redneck rubes, those "less well-educated white people" voted regarding Edwards? Is it just me, or this another example of poor whites being faulted for something the wealthy fucked up? "Obama didn't win? Must be that hillbilly that hauls my junk away." What, if Clinton hadn't won, would it be because the "less well-educated" are sexist?
So, if the poor or less formally educated are so powerful, as this report suggests, why didn't the middle-aged white guy with the populist message win?
Gimme a break. Sounds more like the Dems are already looking for scapegoats if we lose the next election, too. If not, perhaps we'd be more concerned about fixing the patriarchal, racist school systems (designed not by poor whites but by people in the upper classes) then blaming our failures on the people whose primary educational experience is through those systems.

Posted by pbaitch | January 10, 2008 9:35 AM
8

thats wierd
because one of the reasons I think obama would win would be because all these rural based truckers,farmers,carpentors,plumbers and electritions I met hitchhikeing the country last summer...who traditionaly vote republican...and hate the clintons, they love obama out of all canadetes and consestantly comment about how "he's the only decent one of the swine". and of course about half of these people use the word "nigger' in casual conversation.
if rural republican (well independent) racists would vote for obama then anyone could

maybe I'm wrong....maybe only those with enough cojones to reach out and pick of a hitchhiker have the balls to reach across the political/racial divide

Posted by linus | January 10, 2008 9:50 AM
9

Maybe it's because the votes were counted by Diebold machines, noted for their ability to swing elections?

New Hampshire -"Live free or Die-bold".

Posted by sceptic | January 10, 2008 9:52 AM
10

You know, this whole liberal punditry meme about whites and racist attitudes sounds good and all...

but I was under the impression that the polls were fairly accurate regarding Obama's numbers and that it appears simply more people turned out for Clinton than were expected.

Isn't that a flaw in survey methodology then, which is going to be flawed anyway, since it always looks to past turnout?

Posted by Mickymse | January 10, 2008 9:53 AM
11

OK, after all the complaints about what Jesse Jackson Jr. said yesterday about Hillary, what will those folks make of this?

Posted by tsm | January 10, 2008 10:00 AM
12

After all, there's nothing a white man with a penny hates more than a nigger with a nickel.

Posted by Not okay | January 10, 2008 10:01 AM
13

All the pundits were so juiced up on the "Wicked Witch is Dead!", they ignored the huge undecided/not sure group. Obama's numbers were fairly accurate.

Another theory to throw out, some women can't answer "Hillary Clinton" to a telephone poll in front of their husband, then vote for her in the private.

Posted by anna | January 10, 2008 10:07 AM
14

I think a much more probable cause might be:

A. People refused to answer phone calls and surveys in the last week, out of sheer fatigue. You can see this in many anecdotes from on-the-ground reporters.

B. Media spin doctors tried to push it as a probable loss for Sen Clinton, even though they said she would win by 10 points less than two weeks before, and when they got the response they wanted, they stopped polling - no polls were done on Sunday or Monday, even though the election was on Tuesday, while the local cable channels supposedly showed Sen Clinton's breakdown moment and debate performance something like 10,000 times between Saturday and Tuesday.

The media wanted to declare Sen Clinton a winner, but the only way to do that was pretend she was going to lose.

Posted by Will in Seattle | January 10, 2008 10:17 AM
15

Eli,
Karl Rove actually says the same thing in his op-ed in the Wall Street Journal today.

http://online.wsj.com/article/SB119992615845679531.html

It was fascinating to read his analysis, and when I found myself agreeing with him, I threw up a little in my mouth.

Why is he giving Hillary advice on how to win? Is he for sale to the highest bidder now, or does he still think that if HRC is the Dem nominee, that is the best scenario for another Republican win in November?

The Horror....the horror!

Posted by Steve | January 10, 2008 10:18 AM
16

Perhaps polls simply aren't as accurate as pollsters would like us to believe?

Posted by NapoleonXIV | January 10, 2008 10:33 AM
17

Maybe the people who voted for HC just thought she would be a better president.

Posted by crazycatguy | January 10, 2008 10:38 AM
18

@9 - there are indeed a couple of slog threads out there claiming hacked Diebold machines are behind this. Seems plausible. The smoking gun is a bizarre disparity in the results between hand counted precincts (where Obama won by margins close to those polled) and machine counted precincts where Clinton wildly outperformed Obama with total vote counts that were far above poll predictions. At least that's what's being claimed.

Posted by kinaidos | January 10, 2008 11:15 AM
19

as above, the polling was wrong about guessing the winner, but almost all of them got Obama's percentage of the vote almost exactly right. They just didn't account for undecideds.

Who knows why. Maybe some people temporarily switched from Clinton to Undecided after seeing the Iowa results?


Posted by josh | January 10, 2008 11:20 AM
20

Not entirely true with the education gap when you break it down. Clinton won amongst all groups (although "some college" and "college graduate" were too close to be statistically significant) except those with postgraduate degrees, which Obama won handily (43% to Clinton's 31%), skewing the results when considering only college grads vs. those without degrees. The More You Know*

Posted by it's ME | January 10, 2008 11:41 AM
21

So you defend your knee jerk to the small sample size of an isolated vote with... another knee jerk reaction to the small sample size of said vote?

This is why you write for a local weekly.

Posted by Gomez | January 10, 2008 12:16 PM
22

10 and 19 are right. The polls accurately predicted the Obama percentage of the vote, so the whole racist theory is debunked (as is the Diebold theory).

What the polls missed were Hillary Clinton's percentages. So the relevant question is why Clinton got the undecideds and some of the Edwards voters. I think it's pretty easy to conclude that the support for Edwards was soft and that the anti-Clinton media coverage and cheap shots from the Edwards and Obama campaigns provoked a backlash among those voters. But that's not as inflammatory of a theory and doesn't make Clinton look like a racist bitch, so everyone is just going to ignore it.

Posted by Cascadian | January 10, 2008 12:37 PM
23

I just have to shake my head.

Like some of the other posters, I just don't see what the big deal is. The stats were wrong. Big deal. People trust stats too much as it is, because they are too easily manipulated.

Posted by Toby | January 10, 2008 12:50 PM

Comments Closed

In order to combat spam, we are no longer accepting comments on this post (or any post more than 14 days old).