How Bad Were the Polls Last Night?

By: Lowell
Published On: 1/9/2008 7:19:33 AM

On the Democratic side at least, the opinion polls were a miserable failure -- really a complete debacle -- last night.  How bad were they?  The actual results were Hillary Clinton with 39%, Barack Obama with 36%, and John Edwards with 17%. Here are the polls, from best to worst

McClatchy-MSNBC: Obama 33%, Clinton 31%, Edwards 17%
Suffolk/WHDH: Obama 39%, Clinton 34%, Edwards 15%
Rasmussen: Obama 37%, Clinton 30%, Edwards 19%
CBS: Obama 35%, Clinton 28%, Edwards 19%
CNN/WMUR: Obama 39%, Clinton 30%, Edwards 16%
American Research Group: Obama 40%, Clinton 31%, Edwards 20%
Zogby: Obama 42%, Clinton 29%, Edwards 17%

McClatchy got the order of finish wrong, but didn't do too badly (a dead heat within the margin of error).  The rest completely sucked, with Zogby, American Research Group and CNN/WMUR being wildly off the mark.  So much for polling? Maybe it's time for all of us to deemphasize the "horse race" and focus on substance?  Would that be so horrible?

P.S.  I agree with Richmond Democrat: "The pollsters have just revealed themselves to be basically worthless." Actually, I'm starting to think they're downright harmful, not just worthless, to the extent that they skew our politics into an endless horserace and away from substantive discussion of issues. It's also the media's fault -- and yes, the blogs too -- for placing too much emphasis on these things. It's time to change that.


Comments



In other news, Obama wins (Lowell - 1/9/2008 7:21:36 AM)
the endorsement of the SEIU Nevada.  Also,

..Obama was expecting to get another boost from labor in Nevada. The 60,000 member Culinary Workers Union, Local 226 was scheduled to announce its endorsement Wednesday.


Issues would be nice (afausser - 1/9/2008 7:23:31 AM)
That might make people vote for who they actually think is the best candidate :)


Issues? Qualifications? The best candidate? (Dianne - 1/9/2008 8:49:53 PM)
Oh please, we'd rather jump on a bandwagon as soon as it's out of the gate.  Never mind that we don't know where it is going.... we're going for a "uniter" again (not a divider supposedly).

Heaven forbid we'd ruin the "rip" by talking about anything so boring as the issues!  



A constant problem here in NH . . . (Bernie Quigley - 1/9/2008 7:51:56 AM)
A constant problem here in NH comes from the recent movement of people across the border from Massachusetts to the industrial cities in the south of New Hampshire. They are very conservative and in my opinion unimaginative voters. They come from industrial Massachusetts and are representative of working class people who have risen a little ("upper working class") and are terrified of change.

If you took away those southern counties which I have said should be renamed North Massachusetts, Obama would have won in a landslide. Same in '04 - Wes Clark was way ahead in the northern regions and so was Howard Dean, then the south counties voted for Kerry. I really don't see these dynamics so fully expressed elsewhere other than the Northeast and particularly in the extended Boston region. In fact, I have opposed the New Hampshire primary altogether because it only today gives a picture of people of post-industrial Massachusetts and how they would vote. Frankly, Iowa's heartland farmers seem to be less afraid of a black man from Chicago than my cousins at the bottom of the state holding signs which read "Irish for McCain" all over national television these past few days. I was raised up here with South Boston people but spend half of my life in Virginia and North Carolina.

Here is a quote from an article by an excellent local reporter, John P. Gregg giving analysis on last night's vote:

"Sullivan County (bottom of the state - bq) County Treasurer Cynthia Sweeney, who is in her mid-60s, said she voted for Clinton in part because the younthful Obama "scares the daylights of of me" and reminded her of "those same young Turks" (Islamic slur? - bq) Sweeney blamed for the decline of her former employer . . . "It's the attitude of 'we know, we can do it better, get out of our way and let us do it.'" said Sweeney . . ."

This is the politics of Fear and Anti-Hope. I did not find that at in North Carolina or Virginia because there was no such vast industrial class sent upward in only two generations. People outside forget, but - particularly in the industrial cities of northern New England - time has left us behind and left the South and the Southwest to rise.

I's feeling two things at this point: one; The South is still a rising boat and rising boats don't express this kind of fear; the South and the newly empowered economic regions will rise ot Hope and not succumb to Fear and they will rise to Obama and Edwards; two, Mike Bloomberg, who appeared with Sam Nunn and David Boren in a "post partisan" conference in Oklahoma left quickly when he saw Obama's flame ignite in Iowa. He saw his gate close. He saw the gate open again last night as Bloomberg (with his best bud Arnold) is all about being anti-thesis to Billary's Politics without Passion.



Do you think there was any evidence (Lowell - 1/9/2008 7:55:16 AM)
of a "Wilder effect?"  I sure hope that wasn't the case, but it's a well-known "phenomenon which has led to inaccurate voter opinion polls in some American political campaigns between a white candidate and a non-white candidate."  Again, I hope not, but I'm curious what you think.


I don't think so. (Bernie Quigley - 1/9/2008 8:16:18 AM)
It is hard to gauge NH by outsider reporters. We are like two states; the mountains and the industrial cities at the bottom. I am more looking forward to South Carolina and the South because I think that one of the things that happened in Iowa is that black people had pledged to Hillary in loyalty to the Clintons prior to Iowa thinking that Obama didn't really have a chance. When he won Iowa it was clear that he did have a chance and he does. So Obama's support went up 20% in South Carolina (I think I read those statistics here at RK) among blacks.  Gwen Ifill mentioned yesterday that South Carolina is 50% black; Iowa should still bring it to South Carolina in spite of New Hanpshire. There is growing here in NH and I think everywhere a stark generational contrast between Bill Clinton - he no longer draws crowds - and Obama (they faint). As Gregg says, we actually LOVE Obama and Michelle now in the northern valleys. I think that is correct (in the mountains but not in the southern cities). The Clinton supporters up here today all use the same phrasing - I heard this again and again on the C Span reports -  they want to bring back the past because,  it was " . . . better then."  


I don't really think so either. (UVAHoo - 1/9/2008 11:45:43 AM)
One of the very interesting things about these polls is that they wildly miscalculated Clinton's support, but for the most part correctly gauged support for Obama and Edwards (at least within the margin of error).  To me, that seems to point to undecideds breaking disproportionately for Clinton at the last minute.  I think its also possible that Independents didn't vote for Obama at the rate that was widely anticipated.  My understanding was that his poll numbers were buoyed by that assumption.

I also have to wonder about a response bias within the sample.  This is purely speculative, but Obama supporters were "fired up, ready to go" after Iowa whereas Clinton supporters were more demoralized.  That could have translated into Obama supporters being much more willing to participate in polls and voice their support for Obama while Clinton supporters were less responsive.  Just speculation, of course.



I don't think so either (Hugo Estrada - 1/9/2008 11:48:31 AM)
First, we are talking about Democrats. Sure, they can be racists as well. However, the favored candidate here was a woman, which is a pretty progressive choice as well.

Also, why would any Hillary supporter be ashamed of supporting Hillary?  



There was a lot of talk about this (Lowell - 1/9/2008 11:51:52 AM)
on Diane Rehm this morning. The feeling among the political analysts was that it was DEFINITELY a possibility.  Check out the show when it's online...


Bah. (Silence Dogood - 1/9/2008 3:46:53 PM)
The political analysts are looking for convenient means to explain why they were as bass-ackwards wrong as the pollsters for discounting Hillary's chances.  Thinking about it less like a defensive political analyst and more like a regular voter, though, I'd think folks who were disinclined to vote for one candidate because he was black aren't therefore more likely to vote for another candidate who is a woman.  Nor do I think someone who wants to vote for a woman is going to lie to a pollster and say he's voting for a black man because otherwise he might be accused of being a racist--at least not any more than Obama supporters would have lied and said they were voting for Hillary so we didn't think they were sexist.  Maybe the couple hundred people who voted for Biden or Dodd were voting against the black man, but that's not why Hillary overperformed the polls; don't you think it's more likely that the closet bigots, once they got into the privacy of the polling booth, pulled the lever for Southern white male pretty boy John Edwards?

And the Obama/Edwards margin wasn't even close.

A rational attempt to use the "Wilder Effect" to explain Hillary's significant margin--not just over Obama but over the field at large--simply does not hold water.



The debate goes on (Debby - 1/9/2008 8:01:52 AM)
Hillary's win last night means that the American voters outside of Iowa and NH have an opportunity to exercise their right to vote and fel it might matter in deciding the nominee.  The debate will continue at least through Feb 5 and that will only serve to give us all an opportunity to hear more from all the candidates-hopefully an in-depth disucssion of the issues that our country's next leader needs to grapple with and how they intend to do it.  None of this is a bad thing.  If we are all glad to see the energy of increased involvement and turnout through Iowa and NH, then its a good thing for democracy to see that energy continue for at least another month.  Many are tired of hasty decisions under such auspicious circumstances; of never having their votes count in the selection.  Even in 2004, the race did continue through some of the races past IA and NH but from early on, John Kerry was the presumed winner and nominee, and that didn't turn out so well.  American's need and deserve a real opportunity to put these candidates through their paces in order to be sure that the person we're throwing out there to go against the "other side" will come out a winner.  If we don't do it, you can be sure they will.


The polls . . . (JPTERP - 1/9/2008 9:31:13 AM)
may very well have been right.  All they do is capture a snapshot in time.  In a dynamic race a lot can happen in 24 hours.

Either way -- onto Nevada and South Carolina.



What happened in the last 24 hours (Lowell - 1/9/2008 9:32:32 AM)
that could possibly explain a swing from a 10-point Obama lead to a 3-point Obama loss?  


Begin the Beguine (demdiva - 1/9/2008 10:56:11 AM)
People woke up and realized that all the stirring speeches before swooning, shiny-happy people won't get a health care bill passed in Congress, restart the economy, or End The War.  Hillary is battle-tested and tough enough to withstand the GOP onslaught.  What could the haters say about her that hasn't already been said?

The NH results should put the MSM and blogosphere on notice that we are not going to tolerate their grotesque and blatant sexism, but I'm not holding my breath.  It will be a fight the whole way, as well it should be.  The good news is 57% of us are used to fighting, backwards and in high heels.            



"What Happened in the Last 24 Hours... (BP - 1/9/2008 12:49:18 PM)
...that could possibly explain a swing from a 10-point Obama lead to a 3-point Obama loss?"

A front page endorsement of Obama on Raising Kaine?  Is there a Reverse RK Effect at play?

Just kidding, of course.  

(Although we could "scientifically" test for the hypothetical Reverse RK Effect by having someone post an endorsement of Hillary the day before the Nevada caucuses to see if it results in a surprising, come from behind victory for Edwards).



COMMENT HIDDEN (Rebecca - 1/9/2008 9:37:39 AM)


Do you have any evidence of fraud? (Lowell - 1/9/2008 9:39:34 AM)
No, didn't think so.  Every time your candidate loses doesn't mean it was fraud.  There are several other possible explanations, why jump to that one?


Polling in NH based on flawed models (TurnPWBlue - 1/9/2008 10:52:32 AM)
The polling used in NH was based on flawed models that assumed turnout ratios, demographics, trends, etc., similar to the last primary in NH.  By all accounts, voter turnout yesterday looked nothing like 2004, and, thus, the models were flat out wrong.

When they poll, they look at a statistical sample of 1000 or so voters.  They then use historical models to profile how those 1000 sampled voters fit the demographics, etc., from the previous race and the model for this year's race.  in an oversimplified example, If the 2004 race showed that 40% of the voters were white women, they extrapolate how the vote would fall based on the sampled voters who were white women and an estimate of 40% of the turnout falling into this category then do this across their demographic breakdowns.  Yesterday, there were lots of independents and first-time primary voters who are going to throw off the model.  



COMMENT HIDDEN (Rebecca - 1/9/2008 9:45:28 AM)


Screaming "fraud" every time an (Lowell - 1/9/2008 9:58:25 AM)
election doesn't come out your way?  It's you right to do that, but why would anyone listen?


If Hillary had been predicted to win by 10 points.. (Rebecca - 1/9/2008 11:59:09 AM)
If Hillary had been polled at 10+ ahead of Obama and Obama had come in 13 ahead of the prediction I would say the same thing.

I am amazed at the lengths people will to so as not to confront the possibility of a compromised democracy. Its like it only can happen in other countries. It reminds me of the phrase "It can't happen here." As long as people think that way it will happen here more and more often.

I wonder if Americans really would fight the American Revolution if they needed to in these times.



All I have to say is (Lowell - 1/9/2008 12:17:16 PM)


Because she is making a good point, Lowell (Hugo Estrada - 1/10/2008 4:09:05 AM)
Again, I don't think that they were rigged in N.H. Hillary seemed pretty surprised that she won. Normally, when elections are truly rigged, the "winner" is overly confident.

That being said, we should all be more critical about wide fluctuations in polling and results. This should not happen as often as it does. It begs the question about the reliability of the elections.

Back in 2004, I remember that many liberals on Daily Kos dismissed the allegations of fraud in Ohio, saying that Kerry lost fairly. Now, a few criminal convictions after, we know that there was electoral fraud.

The conventional wisdom is that this can't happen in the U.S. The evidence seems to contradict this wisdom.



Follow the leader? (Eric - 1/9/2008 10:00:01 AM)
Despite all the statistical modeling and predictive analysis in hand, it's possible the powers that be (both polling and media) still had grave concerns about releasing a prediction that wasn't the same as the others.  I've personally seen this happen before in other, not to be named, organizations.  Plus, if they were wrong all of them would crash and burn where the reverse, if one group had published a Clinton victory and she didn't win, they'd look like fools and would never be trusted again.

Or maybe their polling methods were off (probably the wrong balance in their sample group).  But that sounds odd because they'd all have to have had the wrong the sample groups or all made similar statistical mistakes to come up with similar wrong results.



"Iron My Shirt" (Quizzical - 1/9/2008 11:23:48 AM)
Did you see the incident the day before the New Hampshire vote, where Clinton was heckled by a guy who got up and pranced in front of the stage with a sign that said, "IRON MY SHIRT"?

She handled that pretty well, and asked if there were any other men in the audience who needed lessons on how to iron their own shirts.

Something happened in New Hampshire to awaken the feminist vote -- maybe that was it.



"Something happened in New Hampshire to awaken the feminist vote" (Silence Dogood - 1/9/2008 3:58:03 PM)
BINGO.

I don't believe Hillary Clinton's teary-moment moved any votes, but I do believe that the reaction of the mainstream media provoked some vote shifting.  As soon as Hillary was misogynistically painted by pundits, analysts, and Obama or Edwards supporters as a weak woman, as soon people started delighted in instituting a Hillary Clinton death watch, as soon as it looked like she was going to get obliterated and that no other woman would ever get this close to the presidency again, the feminists made up their mind to turn out for her.

I want to take a moment to spell this out for some of the dumber Obama supporters out there who are angry that Hillary's still in the race because you dislike her and use words to describe her like "calculating" and "shrill" or even "bitch":

THIS IS AT LEAST IN PART YOUR FAULT.  YOU FAILED TO REALIZE THAT HOW YOU WIN WILL COME TO DETERMINE HOW YOU LOSE.



You hit the nail on the head! (AnonymousIsAWoman - 1/9/2008 11:32:11 PM)
It has even irritated me into posting three articles to my own blog defending her and I'm actually vacillating between supporting Obama or Edwards.  I never was in the Hillary camp but she has been the target of sexism and it is off putting.

Thank you for pointing out that this is causing a boomerang effect that is helping her among women, many of whom have themselves experienced discrimation in their lives.  That is especially true of baby boomer women.  When we were growing up, the world was far different from the way it is now.

There was more actual discrimination against women and it was legal.  

As an example, when former Supreme Court Justice Sandra Day O'Connor graduated at the top of her law school class, the only job she was offered was that of legal secretary.

I just had dinner tonight with a friend who is older than me who experienced much the same discrimination when she first graduated from law school.

Women were not only routinely paid less than men, but it was considered just and fair.  People thought that women would marry and be supported by husbands and were only working to supplement the family income or to be able to afford luxury items.

Yet many women were breadwinners too.  They were widows and divorcees just as today.  

And while today, women still make less than men for doing the same jobs, nobody considers it fair any more.  People know it's discriminatory.

The women who experienced sexism and fought it will rally to Hillary especially when they think that she is being treated more harshly than the male candidates because she is a woman.



I'd love to see post election polls (Rebecca - 1/9/2008 12:01:02 PM)
I think the polling companies should start having post- election polls as well as pre-election polls. That would be very interesting...


Clinton's poll said.. (Rebecca - 1/9/2008 12:08:42 PM)
"Clinton's internal polls had Obama winning by 11%."

Looks like even the Clinton campaign organization's polls showed Obama ahead by more than 3 times the margin of error. Anybody willing to explain that? Be my guest. I'd like to get the bottom of this.



Here's the Iron My Shirt video (Quizzical - 1/9/2008 1:05:35 PM)
http://www.youtube.com/watch?v...


A better one (Quizzical - 1/9/2008 1:27:28 PM)
This Iron My Shirt video has better audio
http://www.youtube.com/watch?v...


Maybe it is hard to poll independents (Hugo Estrada - 1/10/2008 4:16:22 AM)
It seems that independents were largely responsible for Obama winning in Iowa, and it may have also been the reason why they may have voted for Clinton, or for McCain, as my wife told me today. It also seems hard to get a sense on how they are going to vote.

My wife was saying that maybe independent voters leaning for Obama may have voted for McCain instead at the last minute as a tactical move, to guarantee that the less bad Republican candidate will win the nomination.

Was there a slightly higher percentage for McCain than the polls predicted?



Humbuck (Gordie - 1/9/2008 7:59:43 PM)
The polls were 2/3 right, they just issed 9 percent of the voters.

Most of the polls before the election had:

Clinton 30%

Obama 37%

Edwards 19%

How did it end up

Clinton 39%

Obama 37%

Edwards 17%

Okay Clinton got 2% of Edwards, but where did the other 7% come from? Undecideds? I doubt that since there was no mention of undecided's in the polls.

I am dropping out, going to JJ diner Feb. 9th and hoping to hear Clinton, Edwards and Obama, then waiting for Feb. 12th.



Quizzical (AnonymousIsAWoman - 1/9/2008 11:44:28 PM)
Thanks for the videos. In fairness, I don't think that heckler was a Democrat or from any of Clinton's opponent's campaigns.

Unfortunately, hecklers and nuts come from across the board. The sexism that concerns me more is the unconcious kind that is coming from the mainstream media and some normally progressive bloggers and commenters who would never dream of discriminating against any woman they know.  

They are simply anti-Clinton because they perceive her as the Establishment candidate.  She is the most centrist, which is why I don't support her.  But they need to make their case without resorting to what I think of as accidental sexism.

By that I mean by referring to Hillary as shrill, a bitch, an ice maiden, or even by refusing to acknowledge that she once really was an agent of change.  It's just that times have changed and there are other more progressive voices.

Also, Rebecca, with all due respect, I don't think the New Hampshire election was stolen.  I know you mean well but would you at least consider the possibility tht the polls were actually wrong.

They are not all as rigorous and unfailing as you think.  There are flaws in some of them.  For example, they rely on people with landline phones and often don't capture those who exclusively use cell phones.  The way you phrase a question can determine an answer.  And sometimes people just lie to pollsters.

I don't know if any of these are the causes of the polls blowing it.  But I've seen in several election cycles various polls being all over the territory.  I'm beginning to think they really do need to take a look at their product and check for the flaws.  The polls are just getting less accurate and we need to find out why.



A fascinating note on the post election polls (tx2vadem - 1/10/2008 1:09:17 AM)
I saw this on the News Hour tonight.  They said that people who have graduate degrees and make more than $100k a year broke for Obama.  And people who made less than $50k and have only H.S diploma broke for Clinton.  

So Obama wins the limousine liberals and Clinton wins the the lower half of Edward's two Americas.  Amazing!  I'd like to see how that plays out in other states.



Precedent for polling errors in NH (Quizzical - 1/10/2008 1:20:16 AM)
Bob Somerby pointed out today that in the 2000 primaries, the polls had Bush and McCain in a dead heat, and McCain ended up winning by 18 points.  So that kind of thing is known to happen in NH, for reasons that have been discussed already (open primary, small population, strategic voting, weather). http://www.dailyhowler.com/dh0...