How Wrong Were The Polls?

Commenter Brian makes an observation "No one is talking about how the polls actually nailed Obama's number. Obama didn't lose this election. He stayed steady and Hillary surged ahead." That seems to be true. Here's a chart comparing the actual results to the most recent Pollster.com current standard estimate polling average.

Just as Brian says, the difference between the Obama poll level and the Obama vote total level seems to just be your basic statistical variance. The pollsters underestimated Clinton's level of support. People who were undecided as of the last round of polling seem to have gone overwhelmingly in her direction.

[also note the relevance of this to Wilder/Bradley effect speculations]

Comments

Her false mailers make me unsympathetic.

False claims about Obama's pro-choice record:

http://tpmelectioncentral.com/2008/01/ hillary_mailer_hits_obama_on_abortion.php

Rove-style attack politics:

http://www.politico.com/blogs/bensmith/0108/ ProHillary_mail_Be_very_afraid.html

DOVER, N.H. – Facing the prospect of defeat in tomorrow’s primary, Hillary Clinton just made her strongest suggestion yet that the next president may face a terrorist attack – and that she would be the best person to handle it.

She pointed out that the day after Gordon Brown took office as the British prime minister, there was a failed attempt at a double bombing in London and Glasgow.

“I don’t think it was by accident that Al Qaeda decided to test the new prime minister,” she said. “They watch our elections as closely as we do, maybe more closely than some of our fellows citizens do…. Let’s not forget you’re hiring a president not just to do what a candidate says during the election, you want a president to be there when the chips are down.”

http://www.boston.com/news/politics/politicalintelligence/ 2008/01/clinton_heighte.html

Posted by bob | January 9, 2008 12:15 AM

So did some of the support come from Edwards? How can that be when he's a sure shot for the nomination?

Okay, less facetiously, I still want an explanation for the extra Democratic ballots that everyone was reporting needed to be delivered around the state. How is that not totally at odds with the result? Massive women's turnout?

Posted by southpaw | January 9, 2008 12:17 AM

oh please. people howled conspiracy when the bush/kerry numbers were off by less. even if the polls were close on obama, that still doesn't explain how they were so wrong on hillary, particularly with the MSM pushing obama so hard. maybe it's all movement from the last minute hillary tear fest. or maybe the media once again proves it's just not very good at getting things right, has no particular claim to knowledge or authority, etc.

Posted by dj superflat | January 9, 2008 12:18 AM

Folks tend to forget how much of an election gets decided in the last 48 hours, which is after most to all of the polling gets done.

You can have a perfectly accurate poll which is upended by the late surge. Elections tend to break at the end in one direction rather than the other.

Between Hill's tears and Bill's rant, the Clintons owned TV coverage in the last 48 hours. That matters.

Whouley matters too.

Posted by Petey | January 9, 2008 12:19 AM

Am I reading the graph wrong, or did the polls nail Edwards' number just about as well as they did Obama's? If that's so, then the massive bump Clinton must've come from what? All of the other bit-candidates?

Posted by sherifffruitfly | January 9, 2008 12:20 AM

Convential Wisdom was turned on it's head tonight. Sen. Clinton won against all odds, polls, and the pundits' speculation.

She deserves her win. It was hard fought and against every possible pontification, including my own.

http://thepoliticalpost.wordpress.com/2008/01/08/hillary-clinton-wins-nh-against-all-odds-and-conventional-wisdom/

Posted by thepoliticalpost | January 9, 2008 12:20 AM

The polls did in Obama tonight. If there were no polls he would have had a great night with the exact same result. Now it's spun as a great comback for Hillary and, I suspect, there are a lot of really down Obama supporters.

I support all of the major Dem. candidates. I like each of them all for certain reasons. When you get too caught up in the press hype for your candidate this is the kind of thing that happens, especially in NH. Do not put too much emphasis on one contest or another. This is, or should be, a long battle, not a two state coronation. A long battle will help the Dem candidate by forcing them to get better and help the Dem party by keeping the news focused on them.

Posted by bajsa | January 9, 2008 12:22 AM

Obama's internals had him winning by 14%

No secret here: So called moderate republicans will always vote for the republican over the democrat. Obama didn't get the surge he expected because some of his "supporters" were voting for McCain.

Posted by PEI | January 9, 2008 12:24 AM

People seem to be playing down the significance that NH is Clinton town. They turned out the vote for her at the last minute.

Posted by ava | January 9, 2008 12:29 AM

PEI, I presume you understand that your non-secret is completely contradicted by the graph that forms the basis for this post.

Posted by strannix | January 9, 2008 12:32 AM

Thanks for the props, Matthew. I am watching MSNBC and reading blogs and the notion that Obama lost is overwhelming and dead wrong. Also, MSNBC characterizing this as a HISTORIC UPSET is absurd. She was the inevitable candidate 6 days ago. Also, the polling shows Obama won all his traditional areas of college educated, wealthy, etc. Hillary won her demographics.

Figure out why women voted so overwhelmingly for Hillary in the last 24 hours and you figure out how Mark Penn finally got his micro targeting to work. My theory? 1. Contrived Crying. 2. The abortion mailers 3. Turning Hillary into a victim of a Edwards/Obama double team as well as a press punching bag.

Winning in Iowa gave Obama a title fight, which is all he wanted. He needs Edwards to drop out, and he needs to show that he can battle with Hillary. The polls made it seem like he couldn't be beat in New Hampshire, and I think his campaign got ahead of itself, and better have learned a lesson about expectations.

Posted by Brian | January 9, 2008 12:35 AM

Pretty amazing victory - Hillary didn't draw large crowds, she didn't even have the poll numbers - but handily took the New Hampshire Primary. Talk about the unseen guiding hand of fate that always seems to accompany the Clintons. It’s uncanny. Well, I guess it's a real “comeback” after all, a repeat performance, on to the White House: http://theseedsof9-11.com

Posted by Peggy McGilligan | January 9, 2008 12:37 AM

Biased samples

Predictive polling is forced to rely on biased sampling, because only a small fraction of the likely voters whose names are pulled form the hast can be contacted and/or agree to participate. If this small percentage were a random sample, it is true that all we would have to worry about would be the random variance. But it isn't a random sample, people are filtered out or in based on their attitudes towards answering polls, when they're home to answer the phone, etc.

The major pollsters have learned empirically how to tweak their results so that this biased sample, after they add their secret sauce corrections, performs similarly to the actual electorate. But the real electorate in the case of intraparty races, and races involving more than two main competitors, one Republican and one Democrat, is too variable from race to race for them to be able to concoct the right secret sauce corrections to make the biased sample reflect the actual electorate in these contests. Empirically, the polls always do much worse on primaries and races with more than two major contenders, than they do on Dem to Rep, one-on-one, matchups. These particular results are not unexpected or unrepresentative.

Posted by Glen Tomkins | January 9, 2008 12:42 AM

Matt is wrong. Undecideds broke equally for Clinton and and Obama. The polls underestimated the extremely high level of Democratic voter turnout from Clinton's urban base.

Posted by Dan Kervick | January 9, 2008 12:44 AM

Obama outperformed the final DMR poll by winning an additional 6% of the vote. If Obama's internals had him winning by 14 over Clinton it meant they were expecting him to outperform the final AGR poll by 6% of the vote. That 6% was there in Iowa it wasn't in NH.

Posted by PEI | January 9, 2008 12:47 AM

You need another bar on the graph. Undecided. When it came time to actually choose, they chose Hillary Clinton. Matt, this is really not rocket science. All you have to do is check it out. Going into the today the undecided vote was still around 17%.

Posted by ken | January 9, 2008 12:47 AM

@sherifffruitfly

One of the more convincing arguments I've seen for Hillary's support is in the undecided category not explicitly cited in the polls. If you estimate off the graph Matt shows, you'll see that only about 85% of the vote is accounted for in the initial O/C/E numbers. Allowing about 7% for Richardson and the rest, that still leaves about 8% of the undecideds, which is roughly the difference in Hillary's numbers, suggesting that a whopping number of the people who made up their minds late did so for Hillary. Why they did that, well, we'll be seeing a whole lot of conjecture on that in the coming days, but that's the where of it.

Posted by Mercutio42 | January 9, 2008 12:47 AM

One thing I've noticed as well is that Obama and Clinton are tied for delegates at this point. She may have won more votes (which matters narrative-wise) but not more delegates. If this thing does go out into the long-stretch, that's important.

Agreed that Edwards should drop out and endorse Obama if he really wants Clinton to lose. After getting clobbered tonight (by 20% pts, 50,000 votes, and 5 delegates) I don't see his justification for continuing on.

Posted by Jared | January 9, 2008 12:48 AM

Now MSNBC is talking about Race being a factor.

Posted by Brian | January 9, 2008 12:48 AM

I think that many independents, mistakenly assuming that Obama's victory was assured (polls gave him a more comfortable margin than McCain) voted for McCain instead assuming that Obama would win but the republican race would be close.

The polls didn't capture the psychology of independents thinking that one race would be lopsided while the other would be close, and chosing to vote in the perceived close race.

Posted by Tim | January 9, 2008 12:50 AM

"Now MSNBC is talking about Race being a factor."

More specifically, Chuck Todd is saying he finds the Bradley effect plausible.

Posted by Petey | January 9, 2008 12:51 AM

You need another bar on the graph. Undecided. When it came time to actually choose, they chose Hillary Clinton. Matt, this is really not rocket science. All you have to do is check it out. Going into the today the undecided vote was still around 17%.

Exactly. That's what Matt is missing. There are always undecideds in the poll, but not in the actual election, so if you exactly hit your poll average in the election, it means that

when
X = voters who supported you in the polls but bailed at the ballot box
and
Y = undecided voters who broke your way,

then X = Y. Since everybody always gets some of the undecideds, when your poll result = your electino result it means you lost supporters.

Posted by Trickster | January 9, 2008 12:52 AM

Matt,

It's not the pre-election surveys that are the key to see what happened here it's the exit polls.

Clinton's key demographic of the over 40's is more numerous in actual voters than Obama's under 40 and dominantly under 30 core. It's entirely a generational thing, and the Matt/Young Ezra Klein demo just doesn't have the numbers yet to overcome your parents.

Go look at the CNN exit poll for NH, and compare it to the entrance poll for IA. They tell the exact same story.

The two dominant lessons here are that registered democrats, who btw also dominantly voted for Clinton, are hungry for the return of the wonk, and that the Media driven Clinton bashing rather than serving as red meat for "conservatives" is a much much more potent source of energy for Clinton's own base.

I think there are lot of people who were of voting age and older in 90s that are still very very upset with what was done to Hillary and Bill during that presidency by the media. And it's exactly that crowd that is who are older than Obama's "new voter" demographic, that are voting for her, while those who by their youth and non-involvement missed the 90s outrages are voting for him instead. The majority of under 25's, his strongest demographic, essential have no memory of the endless overly negative 90s Clinton Press attack machine. But the older you get the more outraged it makes you, especially to see it happening again in the face of the monster that same media has empowered over the last 7 years.

Long live the wonk! Long live the age of Media discontent!

Posted by smacfarl | January 9, 2008 12:53 AM

As one data point in support of my theory, the Political Lunch video podcast interviewed a guy who said he had planned to do just that (before results were known).

Posted by Tim | January 9, 2008 12:54 AM

"I don't see (Edwards') justification for continuing on."

Bill Clinton lost IA and NH in 1992, y'know.

-----

For all the Obamabots who demand Edwards drop out to save us from Clinton, you should understand that we have a majority nomination process.

Clinton can't win the nomination by getting 40% in every race, whether she's running against one or two or seventeen other candidates.

If she can get up to 51% and stay there, she wins no matter how many other candidates are in the race. If she can't, she loses no matter how many other candidates are in the race.

Posted by Petey | January 9, 2008 12:55 AM

Can we stop the talk about Race being a factor? It is totally unhelpful to the discussion.

Posted by Brian | January 9, 2008 12:57 AM

It is not the fucking Bradley effect. Obama received about the same percentage of vote that the polls said he would. The story is that the Obama surge did not materialize whether because the were voting for McCain or because Clinton did better in the surge demographic (young and independents)

Pundits are dumb.

Posted by jack | January 9, 2008 12:59 AM

Of no significance whatsoever: The Guardian has a pretty unflattering picture of Clinton on their front page.

Posted by esa | January 9, 2008 1:03 AM

If the "undecideds" mostly - or half anyway - went for Clinton, then the question is: why?

I think Petey and Brian have it right. The last minute Clinton attacks and the last minute press coverage of "The Crying Game" and Bill Clinton did it.

Which has to make you sad for the electorate in this country that votes for whoever was in the news most in the last 24 hours.

It's like the TV Guide "Most Beautiful Women on TV" polls - whoever was in a "Movie of the Week" LAST WEEK wins.

It's pathetic. This is how a superpower chooses its leaders.

Bring back the Roman circuses - they made more sense.

Posted by Richard Steven Hack | January 9, 2008 1:03 AM

I, hereby decree, that Obama, in the best interest of America, should drop out of the race so the Democrats do not nominate Clinton.

Being a member of Gen X and all around stand up guy, Obama should know you can't claim n-gun when Edwards already yelled shotgun while kissing a baby and touching a donkey. Obama must drop out because if he goes against the iron law of firsties he can no longer claim the change mantle, is no longer a stand up guy and will lose his soul and beginning listening to corporate pop.

I urge all Obama supporters to contact their candidate and demand he end his campaign. In doing so you are not only helping America but also ensuring Obama keeps his soul. Win-Win

Posted by Firsties | January 9, 2008 1:08 AM

"I think there are lot of people who were of voting age and older in 90s that are still very very upset with what was done to Hillary and Bill during that presidency by the media."

And stoking that resentment worked quite nicely in the 48 hours before NH.

But I think it remains a very open question whether or not that resentment is the basis for a successful nomination campaign that is going to last a minimum of a month.

My take is that re-fighting L'affair Lewinsky works well for Clinton in the short-term, and badly in the long-term.

Posted by Petey | January 9, 2008 1:09 AM

Question: Before Iowa, Obama was behind steadily in the double digits. In 5 days we saw the gap flip. Does that seem odd to anyone else? Can you look at this as Obama closing the double digits to 3%?

Posted by Stellaa | January 9, 2008 1:10 AM

It wasn't the tears (maybe the media coverage of the mean media coverage of the tears), it was the Bill factor.

He came after Obama and people listen to Bill.

Bill, after all, is a popular guy.

Posted by hlah | January 9, 2008 1:15 AM

Yes Stella - Obama won NH just like Edwards won Iowa - by gaining in the polls right up to the point the voting started and he came in second. Unfortunately for Obama, he was suppose to win by 10 points whereas Edwards had not been in front in Iowa since 2006 so his win lacks a little luster.

Posted by claim victory and move on | January 9, 2008 1:15 AM

As a MA native this was about Romney. Indies reacted to the polls, saw O was ahead and figured why not bury Romney, the douchebag. McCain exceeded the margin, and O didn't. O really needed indies because 57% of the Dem party is older women. Yet, despite all that, in a straight raw h2h Obama still beat McCain. The good news out of this is Romney is on the ropes, the bad news is that Hillary is not, and they have gotten smarter, and will manufacture a mandate and call this the end of Obama's campaign--rightly or wrongly. People will be embarrassed for "hoping" and by the time they wake up, it will be too late, Hill will be sitting at 2,015. It is so depressing.

Posted by JC | January 9, 2008 1:22 AM

The poll that was farthest from reality, Zogby (13 point lead for Obama) only polled registered Republican and Democratic voters. But NH's primary is semi-open, all voters can vote and switch sides if they wish. I always vote Dem but I've never been a registered Dem.

Posted by anon | January 9, 2008 1:23 AM

Matt,

For comparison and to either confirm or rule out my theory, how was turnout on the Republican side? More than expected? If so, that would be consistent with my theory (Independents thought the democratic side would be a blowout, so they voted in the republican race instead, thinking it would be closer).

Posted by Tim | January 9, 2008 1:24 AM

Perhaps Clinton's surge was due to female voters deserting Edwards over his badly choosen opinion of her Tears Moment.

Posted by bd | January 9, 2008 1:28 AM

Well there is also the theory Clinton herself hinted at in Iowa, when she said the caucus system was unfair to women. I think she believes there are a lot of women who support her, but whose husbands don't, and that these women are uncomfortable about standing up publicly for their choice against their significant others in an open setting like a caucus room.

When women are polled in New Hampshire, their significant others are often within hearing distance - so maybe they lie or dissemble to the pollster. But then once inside the privacy of the voting booth they vote for their secret fave Hillary.

In any case, it looks like we have a fairly significant gender gap going on.

Posted by Dan Kervick | January 9, 2008 1:34 AM

The polls did in Obama tonight. If there were no polls he would have had a great night with the exact same result. Now it's spun as a great comback for Hillary and, I suspect, there are a lot of really down Obama supporters.

Yep. That and the 'unified narrative' line. Look at December's polling, and you basically see Obama coming up through the 20s. The expectations game of the post-Iowa polling totally changes the perception.

If Iowa and NH happen on the same day, I'd imagine a pretty similar result, but with a very different takeaway. Perhaps you have indies seeing the polls, thinking 'blowout', and voting for McCain instead, but I don't see any big shift in the basic trend lines for the Dem candidates.

Posted by pseudonymous in nc | January 9, 2008 1:36 AM

bd

Do you mean the 1% Edwards underperformed somehow correlates to the 8% Clinton overperformed.

Posted by crack | January 9, 2008 1:44 AM

According to the MSNBC's people, that late undecided vote broke from Obama by 1 point, 37-36.

Posted by MNPundit | January 9, 2008 1:45 AM

I absolutely LOVE what happened tonight ! Hopefully now nobody will take the polls too seriously, especially the ones for which the spreads are within the margins of error.

Posted by anon | January 9, 2008 1:45 AM

There's a rule scientists use that says the simplest explanation is probably the correct one. Forget the Wilder effect, or the Bradley effect. Forget trying to explain the results by Hillary Clinton's tears or Chris Matthews arrogant comments.

The polls were just wrong.

They predicted an 8-12 point win by Obama. They don't get credit for "nailing Obama's number" (in fact the pollster image you linked to shows that the variation in Obama's numbers from poll was larger than Hillary's so pick your poll).

Polls are measures of relative support. To nail one candidate's numbers and miss his opponent's result by 10 points is just to fail completely.

Posted by Jinchi | January 9, 2008 1:48 AM

Oh my god, the stupid pundits on TV are talking about that stupid "Bradley Effect"! This has nothing to do with race. I suspect Hillary supporters were driven to the polls because of Obama's lead in the polls.

Posted by hillary supporter | January 9, 2008 1:50 AM

I am an upper middle class white guy who has serious suspicions about how the polls could have been so wrong. If I am harboring conspiracy theories, how is this going to play in Harlem, on the South Side of Chicago, in much of South Carolina, an in every "ethnic" neighborhood throughout the country?

Read African-American weeklies in days head, if you think I am blowing smoke.

Posted by JGH | January 9, 2008 1:53 AM

I think this was absolutely a backlash at the media’s gleefully dancing on Hillary’s grave, together with a real queasiness about the way Obama was being touted as a sort of pop-messiah (the rock star thing, the Oprah thing, the no tie thing, and the Obama is “The One™” thing). He also seemed (mostly through his boosters) to be a less than gracious winner (the “kick her to the curb” booing, the trash talking, the weirdly orgamisic articles which may have seemed like gloating). I think the combination of anger at the way she was being treated in the media combined with unease about Obama as a political figure made voters not want to kick Hillary while she was down. I don’t think this was a political thing at all. I think that whatever happened tonight happened on a largely nonpolitical, deeply emotional level.

As it happens, John Edwards is very much my preferred candidate and tonight was a serious (maybe fatal) political defeat for him. And I am by no means a Hillary fan. But the attacks on her by the media and even some fellow Democrats were so primitive, so visceral and so outrageous that it was difficult not to feel for her and her family. Which, in a way, transformed tonight from a political choice to a very personal one about feeling someone’s pain and choosing to support her rather than give her tormentors the satisfaction of seeing her humiliated.

There was something very touching, even uplifting about her victory after having been so outrageously vilified and slimed by the media and the Obama people. I think I was reached in a very deeply emotional way that I can’t explain or even begin to describe---I still very strongly support Edwards but I actually feel happy that he lost tonight.

I suspect that many others, including many voters in New Hampshire, were touched in the same way and felt the same need to rally round and support Hillary regardless of whether she was their first choice in this election. And that's what I think did Obama in, not racism.

Posted by Mitch Guthman | January 9, 2008 1:56 AM

Can we talk about how the Huey Long effect?

The MSM, all tv pundits, Barack Obama, bloggers and Democratic primary voters all discount how much America wants populist policies. It is a conspiracy I tell you.

Posted by jonny cornpone | January 9, 2008 1:57 AM

Is someone actually calling Hillary Clinton a populist?

Posted by JGH | January 9, 2008 1:59 AM

I am an upper middle class white guy who has serious suspicions about how the polls could have been so wrong

Seriously people, just look at the numbers. The polls didn't agree with each other even when taken on the same day. Zogby had Obama at 42. Fox had him at 32. Polling isn't an exact science.

There's no conspiracy here.

Posted by Jinchi | January 9, 2008 1:59 AM

After watching the coverage of the New Hampshire results, I am astounded at the lack of insight shown by the commentators on MSNBC and other networks. Why did older women come out in force to vote for Hillary? Because it is time for a woman president, Hillary is qualified, and Hillary has concrete ideas and plans based on experience. Obama, though nice, does not have the experience needed to lead us out of the mess created by the Bush administration. Moreover, his ideas are both vague and not in anyway new. As an older voter, I have heard it all before and Hillary is right, talking about change and making it happen are two very different things.

Posted by Diana Zinn | January 9, 2008 2:00 AM

Jinchi: "To nail one candidate's numbers and miss his opponent's result by 10 points is just to fail completely."

Nobody's denying there was a poll failure. The question is why.

If you only had two candidates, and you nailed one's numbers, by definition you would have nailed the other one as well. The fact that there are multiple candidates ruins that outcome.

So the issue is, who switched from who to Clinton to drive her over the expected outcome?

Did people desert the candidates other than Obama, or was it the undecideds, or did people switch parties, or all three? Or, to cover it all, did a bunch of people appear out of nowhere who were suddenly in favor of Clinton (that's called vote fraud)?

The fact that Edwards numbers were nailed appears to make it clear that it was the undecideds. I doubt a lot of Republicans switched parties to vote for Clinton.

What happened here is a bunch of older white women who were undecided voters came down en mass for Clinton based on 1) "The Crying Game", 2) their emotional reaction to the media criticism of Clinton, and 3) media focus on Clinton over the last 48 hours.

I don't see any other rational explanation.

Posted by Richard Steven Hack | January 9, 2008 2:01 AM

Can we talk about the American Idol effect? If people are making voting decisions in the manner speculated, then they are fucking stupid.

Posted by blah | January 9, 2008 2:02 AM

The tears are Hillary's Checkers speech. That desperate, that brilliant.

A classic, Oscar-worthy performance by a savvy politician who is about to give up on the Presidency after one round.

Obama was so stunned by it that he didn't even bother to counter. By the time he knew it hit him, it was over.

Slick Hilly the Comeback Queen

Posted by Jimmy James | January 9, 2008 2:04 AM

Many people, especially on the left, do not trust voting technology. Reference Florida (2000), Ohio (2004), for examples. The polls have never been this wrong. It angers me, and I am white. But my anger pales in comparison to that of other people. Go to a barbershop in north St. Louis tomorrow, and see what they think. You can call it irrational until the end of time, but there it is.

Posted by JGH | January 9, 2008 2:04 AM

Jinchi: Vis-a-vis the poll variances, well, of course, polls aren't going to be totally accurate between themselves.

We're not talking about that variance. We're talking about how one poll nailed Obama's numbers and Edward's numbers, but Clinton did much better than predicted.

That's an entirely different issue.

Explain that one poll.

Posted by Richard Steven Hack | January 9, 2008 2:05 AM

If there were a Bradley effect, wouldn't we expect to see a discrepancy between exit polls and actual results? They seem pretty close.

Posted by RL | January 9, 2008 2:10 AM

Something else about the graph in the OP:

Is it proportional to total voter turnout or not?

If the polls underestimated total voter turnout, then Obama getting what the polls predicted he'd get doesn't make the polls right about Obama. It just means two errors canceled each other out.

Posted by Tim | January 9, 2008 2:11 AM

Turnout was bigger than ever. Over 500,000

Posted by Jimmy James | January 9, 2008 2:13 AM

Diana, your points are simply Clinton talking points. They're irrelevant in explaining why this happened in NH and not in Iowa.

We know older white women will prefer Clinton over Obama, although not necessarily for the (allegedly) rational reasons you state.

For the record, I don't think any of that is true in Clinton's case in any event: she's not a change agent, she's corrupt, she's a hawk, and she's probably going to find it hard to beat even the loser the Republicans nominate.

Worse, while she may have concrete plans and ideas for domestic issues, right now foreign policy is critical - and she's a disaster waiting to happen on that front. She's an AIPAC today, clueless about Palestine, Iraq, Iran, and Pakistan, and that means she is a real danger to the security - and economy - of this country.

What difference does it make if she has great ideas on health care if she starts a war with Iran, and the oil spike and the Chinese dumping the dollar crash the economy?

(Of course, she's considerably better than Giuliani, but that's not saying much given what a loon he is.)

Posted by Richard Steven Hack | January 9, 2008 2:14 AM

Let's just see who sits out November and lets the McCain waltz in now.

Posted by JGH | January 9, 2008 2:16 AM

s/b "AIPAC toady"...

To expand on the Iran issue, the problem with ALL the candidates (except Ron Paul and Kucinich, who are irrelevant) is that you cannot approach the Iran issue from the standpoint that "Iran is a threat" and "Iran must give up enrichment." That isn't going to happen.

So anyone working on that basis - and Clinton is (and so is Obama) - will end up failing on the diplomacy side, no matter how much they want to use diplomacy instead of the military option. That leaves them at a dead end.

So Clinton may not want a war with Iran (unless she's even more of an AIPAC toady than I think, which is quite possible), but she will still end up with one because her basic premises about the situation are incorrect.

The degree to which the US must use diplomacy with Iran is far greater than any candidate has indicated they are willing to go - and the reason is the Israel Lobby will trash them if they do indicate such a willingness.

Foreign policy, because of its impact on the US economy and security, is THE critical issue of the next decade and probably longer in this country. The US electorate cannot afford to get it wrong in this election, like they did in 2000 and 2004.

And Clinton is not right.

Posted by Richard Steven Hack | January 9, 2008 2:21 AM

I think that it was a combination of a backlash against the mediapredicting prematurely a big Hillary loss and the bradley effect. How else can you explain that the polls were right about McCain winning but wrong with respect to Obama?

Posted by MGJ | January 9, 2008 2:23 AM

Explain that one poll.

I'm not sure which of the 20+ post-Iowa polls you're referring to, but the simplest explanation is akin to firing a shotgun at a target - some of the pellets are bound to fall near the bullseye - that doesn't mean they were aimed any better than the rest.

And as for the Wilder/Bradley effect comments - people are basically arguing that all of the polls were correct: that Obama had a 5-10 point post Iowa bump and then Clinton rallied from behind - all in 5 days. The only way to make that argument is to believe that the polls were measuring dynamic changes in support over a very short time period - despite the fact that many of the polls were done concurrently over several days yet don't agree with one another.

Posted by Jinchi | January 9, 2008 2:29 AM

Someone keeps talking about the independent voters as a reason for Hillary's win.

Unfortunately for their reasoning, analysts say that independent voters broke significantly for Obama. So did young voters.

I think that Hillary's teary moment, which I don't think for a moment was scripted, helped voters (especially older Dem voters who might not have been planning to go to the polls) see that she is a human being and really does care.

I'm for Hillary, but I'm thrilled that the candidates in the running all seem to be made of sterling stuff. Even John Edwards, who I always see as a fast-talking used car salesman, has been impressive. I would be thrilled to have Hillary or Obama running the country, and I think one of them will be. I still think that a Hillary/Obama ticket might be a one-two punch to remember. That or Hillary/Bill, which I think is unlikely. But there are many other great possible VPs out there. All of them better than Darth Vader, aka Cheney.

Posted by Gene Venable | January 9, 2008 2:29 AM

Jinchi: "I'm not sure which of the 20+ post-Iowa polls you're referring to,"

The one at the top of the page: the average of the polls.

If the average is that far off on Clinton, but dead on for Obama and Edwards, where did the Clinton votes come from? That's the crux of the discussion.

You can't hit the average on the nose for Obama and Edwards, and be off by a full 10-11 points on Clinton without something being wrong ABOUT CLINTON.

"independent voters broke significantly for Obama. So did young voters."

Yes, but if older white women UNDECIDED voters broke for Hillary en mass, as well as some formerly Obama older white female voters, Obama might make up the difference in other independents, to keep his numbers, but the white female voters would drive up Clinton's. Supposedly the undecided were 17%, enough to explain an 11-point shift to Clinton while still giving 8 points to Obama.

This is the only rational explanation for the disparity in the average poll numbers.

It's not the disparity in specific polls that matters here, it's the disparity in the average at the top of the page.

Posted by Richard Steven Hack | January 9, 2008 2:51 AM

Some Obama supporters are not gracious in victory or defeat.

Posted by O'Reily's advocate | January 9, 2008 2:55 AM

You can't hit the average on the nose for Obama and Edwards, and be off by a full 10-11 points on Clinton without something being wrong ABOUT CLINTON.

You can if the polls themselves are systematically in error. An average of wrong numbers can still yield a wrong number - it doesn't matter that it was closer to Obama and Edwards.

In the shotgun analogy, that's like the shooter aiming just a bit too far left.

Posted by Jinchi | January 9, 2008 2:58 AM

Blackboxvoting.org is reporting that the hand count ballots went 60/40 in favor of Obama, and the electronic count went 90 percent Hillary.

I do not for one minute believe this election wasn't hacked.

Posted by Jay Esbe | January 9, 2008 3:00 AM

When Obama and Edwards ganged up on Hillary at the debate, when the media piled on her planning her political funeral, when she became emotional, in the mids of the undecided she became the underdog in the truest sense of the word! That is why they all broke towards her at the end.

Posted by Democrat | January 9, 2008 3:00 AM

A question about the “Bradley effect”: Can anyone explain why a white person being asked about his or her preferred candidate in the Democratic primary in New Hampshire would feel socially uncomfortable expressing support for John Edwards, Hillary Clinton, Bill Richardson or Dennis Kucinich and so feel pressure to express false support for Obama to avoid seeming bigoted in the eyes of a pollster?

As I understand it, this is the basis of the “Bradley effect” so in the absence of such an explanation or a plausible theory which accounts for why whites in New Hampshire would feel the need to falsely claim to support Obama, I don’t see how the “Bradley effect” is relevant to the result in New Hampshire

Posted by Mitch Guthman | January 9, 2008 3:04 AM

Notice from 49 yr old black female - the community is not happy w/the Clinton's. You don't say the things about a minority the way they did and expect us to vote for Hill. Look for the urban/black talk radio to be buzzing big time tomorrow. It actually started on Tuesday and there was so much excitment for Barack. Bill and Hill are toast to most black Americans. Absolutely toast.

Posted by PBrowne | January 9, 2008 3:11 AM

Some Obama supporters are not gracious in victory or defeat.

Obama supporters are never gracious.

Notice from 49 yr old black female - the community is not happy w/the Clinton's.

Are you a Jonah Goldberg parody?

Posted by Dan the Man | January 9, 2008 3:17 AM

"You can if the polls themselves are systematically in error. An average of wrong numbers can still yield a wrong number - it doesn't matter that it was closer to Obama and Edwards."

Yeah, but now you're claiming that the average of polls at the top of the page is off because ALL the polls were SYSTEMATICALLY off in the same direction.

Now explain how that happened, if the polls mostly didn't agree with each other as you said earlier.

And it does matter if it was closer to 2 out of 3 of the candidates - that again emphasizes that there was something wrong with the CLINTON estimate - OR the Clinton performance.

And I'm saying the difference in the Clinton performance is clearly explained by an obvious demographic: older white women undecided voters (and probably many formerly pro-Obama older white female voters, some of whom have posted here in other threads tonight) influenced by 1) "The Crying Game", 2) Bill Clinton, who they still lust after, 3) their emotions over the alleged press beating Clinton was getting, and 4) the overall media focus on Clinton in the last 48 hours.

Jay: Thanks for the Blackboxvoting.org connection. On their page I see this:

"John Silvestro and his small private business, LHS Associates, has the exclusive programming contracts for all New Hampshire voting machines, which combined will count about 81 percent of the vote tomorrow."

"The exact same make, model and version hacked in the Black Box Voting project in Leon County is used throughout New Hampshire, where about 45 percent of elections administrators hand count paper ballots at the polling place, with the remaining locations all using the Diebold version 1.94w optical scan machine. Because the voting machine locations tend to be urban, this represents about 81 percent of the New Hampshire voters."

They're still not sure if the disparity is some sort of "urban effect" vs the rural towns.

However, this post was interesting:

"My mom, aunt, and dad all voted for RP [Ron Paul] today in my hometown, My mom and aunt both work passing out ballots, and checking them off. I just looked at the politico map and it says their town has ZERO votes for Ron."
http://www.ronpaulforums.com/showthread.php?t=82599&page=18 "

They are starting to call this Clinton "surge" "fishy".

Posted by Richard Steven Hack | January 9, 2008 3:18 AM

To DantheMan - Since I have no idea what Jonah has been saying I can't answer your question. But DantheMan, parody or not, Hill and Bill have really done damage to their reps because of all the negativity towards Barack. Its as if how dare this black man challenge us? We are the ones with the power. The audacity of that black man to think he can challenge the Clinton power and get away with it. It is obvious the Clintons will say and whatever they need to do to get power and the dems are stupid enough to think they should have it. Make fun if you want, but this is serious to many of us and if you are not my kind then you wouldn't be able to begin understanding where I am coming from.

Posted by PBrowne | January 9, 2008 3:26 AM

By community and most black Americans I assume you actually mean you and your friends. This proud black woman will be voting for Hillary on February 5th.

Posted by Q. | January 9, 2008 3:27 AM

Has anyone considered that it was a really, really nice winter day and that more old people got out and voted mostly for Hillary?

Was it the weather?

Posted by Kid Sunshine | January 9, 2008 3:28 AM

To DantheMan - Since I have no idea what Jonah has been saying I can't answer your question.

Ah, a Jonah Goldberg parody who denies he's one. Even better. Congratulations!

Posted by Dan the Man | January 9, 2008 3:30 AM

Here's an interesting question.

What happens to the Clinton campaign if the "Clinton surge" turns out to be an electronic voting machine error?

What happens to it if it turns out to be vote fraud?

Just asking, since there's no evidence yet aside from a the major disparity between the hand-counted ballots and the electronically counted ballots which might have a reasonable explanation.

Posted by Richard Steven Hack | January 9, 2008 3:30 AM

Q - doesn't matter, you are a true minority for Hill, but good for you. I hope she makes you feel proud to be a woman and an African-American. She and Bill sure said a mouthful in saying what they thought about Barack. Guess that makes you proud. Again, good for you.

Posted by PBrowne | January 9, 2008 3:30 AM

RichardSHack - I bet we'll never know. Bill and Hill have lots of stroke and sometimes you just can't beat that kind of power. The only solace for me is that it was close and they both have the same delegate count.

Posted by PBrown3 | January 9, 2008 3:34 AM

When Obama and Edwards ganged up on Hillary at the debate, when the media piled on her planning her political funeral, when she became emotional, in the mids of the undecided she became the underdog in the truest sense of the word! That is why they all broke towards her at the end.

And yet, the piling on was by all accounts much more severe on the Republican side, where McCain and Huckabee beat up on Romney mercilessly in the GOP debate. No sympathy for Romney obviously, or backlash against McCain.

If you're in the Obama and Edwards camps, this has to give you pause, when the sort of par-for-the-course treatment that would go without mention in an all-male setting is apparently so offensive to a significant percentage of Democratic women as to change the way they vote. The run for the handkerchief might work in a Democratic primary setting, but it won't in the fall.

Posted by ChuckE | January 9, 2008 3:53 AM

I fear that the story is going to become Obama's Huge Collapse, which sucks, because it demeans both he and Clinton: it ignores that Obama, as you note, performed as well as was expected (and that he really did come damn close to winning), and it minimizes what was a really, really great electoral performance by Clinton.

Posted by Jon | January 9, 2008 4:05 AM

ChuckE, remember how many Republicans support torture? They have a lot of sadistic bastards on their side.

Posted by MNPundit | January 9, 2008 4:17 AM

The rise of "you go girl" politics. Awesome.

Posted by blah | January 9, 2008 6:07 AM

The story is the real losers are the pundits and the pollsters. The winners are the people of America and its democracy.

The pundits and pollsters should shut the fuck up and the "journalists" should report the results of the voters and stop trying to drive the process.

Posted by bajsa | January 9, 2008 6:24 AM

Nothing much to do with the topic, but can someone tell me why Wyoming doesn't count to anyone but the guy who won it and who comes off as desperate even bringing it up?

Posted by aleks | January 9, 2008 6:30 AM

Gee ??? The Polls got it wrong ??? OMG !!!!

I was just wondering 'why' the polls and pundits all 'decided' that Iowa was the only primary that mattered. All the rest are doggie poo.

Personally, I'd like to see Edwards, Richardson and Kucinich up there with Obama and Clinton; it'd be one hell of a convention!

Posted by Tahut | January 9, 2008 7:10 AM

I send condolences to Obama supporters. Congratulations to Hillary supporters. Some have suggested Edwards should get out of the way to support Obama. Since I am supporting Edwards, and supported him in the last electoral cycle as well, perhaps supporters in both camps will now have have an inkling how we Edwards folks felt coming out of Iowa twice in four years. Why do we still hang in there even though he hasn't won one of the first two contests? Edwards himself, of course, is imperfect. I realize that after seeing him lose. I can even understand some feeling that he reminds them of a used car salesman. But he and Elizabeth also have experienced and survived real pain in their lives, and although we are perhaps a bit too familiar with it by now, he did come from common roots. While we recognize his human flaws, I think we are impressed by his determination, his tenacity, his consistently progressive policies (a much better healthcare plan than that proposed by Obama) and his willingness to fight for them and for the ordinarly working people they would benefit. He continues to advocate for them with dogged strength and an optimistic demeanor. We are convinced that there are serious flaws in both his competitors (Hillary's high negatives and ability to galvanize the opposition, coupled with the public's desire to move beyond the Clinton/Bush presidential cycles; Obama's lofty rhetoric but less substantive policies and lack of experience). We think Edwards' strengths (years of both real-life experience as a top advocate taking on and beating powerful banks of corporate lawyers; rural working class roots; win against the Helms machine in NC, ongoing support for labor etc.) bode well for him to be the best possible winning nationwide candidate. Regarding questions about where Edwards supporters go now, who knows? Our bottom line is a democratic victory (because we feel that it will forward the policies we favor), but because we honestly feel that the party is missing its chance in that regard by focusing on only the two frontrunners. I think a number of us Edwards supporters will think long and hard about which of the less effective candidates will have the best chance in the general election. Speaking for myself, if Edwards drops out I may just stand back and watch the process until the winner is selected. Hope he keeps fighting for the progressive economic policies he represents.

Posted by popudem | January 9, 2008 7:30 AM

Wouldn't the Bradley effect apply to exit polls as well? Since the final numbers ended up close to the exit polls I think we can discount that factor.

Posted by Charlie | January 9, 2008 7:56 AM

Wouldn't the Bradley effect apply to exit polls as well? Since the final numbers ended up close to the exit polls I think we can discount that factor.

Posted by Charlie | January 9, 2008 7:56 AM

We have been complaining for years about the "kewl-kids", personality-driven, celebrity-tabloid gossip journalism with which folks like Chris Matthews and Tim Russert have projected their personal prejudices into the political process. Lots of voters have figured out the same thing.

Now, we see voter disgust at our journalism moving votes. I think that some of the late surge to Hillary is exactly what we have been wishing, and waiting, and hoping and praying for -- voter pushback against the media. Long Live Bob Somerby !

While this has benefited Hillary right now, as a long term development, it is great news for all progressives.

Posted by Tom in Ma | January 9, 2008 8:31 AM

Glen Tomkins | January 9, 2008 12:42 AM
Predictive polling is forced to rely on biased sampling, because only a small fraction of the likely voters whose names are pulled form the hast can be contacted and/or agree to participate. If this small percentage were a random sample, it is true that all we would have to worry about would be the random variance. But it isn't a random sample, people are filtered out or in based on their attitudes towards answering polls, when they're home to answer the phone, etc.

Bingo!

Random sampling only works reliably on events that are periodic or that occur according to a random probability distribution!

If every voter determined their choice by rolling a dice, or picking the candidate round robin according to their order entering the voting building, the polling would work perfectly!

Polling is a psychological one and the idea that you can somehow magically use random sampling for non-random nor nor-periodic events, is the problem.

Posted b