Gwen, starting with you, do you have a “wrong” theory to share with us?
GWEN IFILL: I have a “reported wrong” theory.
JIM LEHRER: OK, all right.
GWEN IFILL: There’s five pieces to the wrong theory, OK?
There’s one, and it’s from the Obama point of view, I’ll tell you, because that’s — I was told this by an adviser to that campaign today. One is they think that there was an Edwards drain, that is that people went and voted for John Edwards.
Now, the polls show that actually — that the polls were correct when it said how many voters that he would get, Obama would get with the increased turnout, but, you know, they still say other people went away.
McCain drain. A lot of independents who they would have expected to vote for Barack Obama, who did in Iowa, instead voted — inexplicably, as Stuart was pointing is out — even though they may not agree with a lot of the same things that John McCain stands for, they went to vote for him because he seemed independent.
The women drain. There was a lot of discussion — and this was an Edwards factor, too — that in Edwards siding with Barack Obama at that debate and taking a little shot at Senator Clinton, that women said, “Hey, don’t pile up.”
And then, of course, there was — as I keep calling it — the weepy moment, so that there was women sympathy drain.
Then, campaign advisers really believe that their lead was depressed because of all these polling numbers showing that there were double-digit leads and that a lot of people who would have otherwise supported Obama said, “Oh, he doesn’t really need me. He’s fine. Let me look at somebody else.” Because basically most Democrats like all the candidates.
And the fifth is that college students in Durham, the University of New Hampshire, and in Hanover and Dartmouth College didn’t turn out in the numbers they had hoped for. And they had come to depend very heavily on college students in that youth vote.
I have a different theory. The polls were not wrong (except maybe the exit polls). The pundits were.
Consider. In the 1989 race for governor of Virginia, Wilder supposed led Coleman by 11 points, but won only narrowly. The polls were declared wrong. But in fact, the last Washington Post poll put Wilder at 52 percent, and his actual vote was 50.2. Pretty close. The poll number for Coleman was 41, but he got 48 percent of the vote. In other words, people who were declaring themselves undecided ended up voting massively for Coleman, not splitting evenly as pollsters and pundits expected.
Race was the obvious factor in that election (the so-called Bradley effect), and I don’t think it was so much here. But again, the actual numbers for Obama given by the last polls were pretty close to what he got. It’s just that the so-called undecideds (which were a very, very large number) broke massively for Clinton, whose numbers had been trending up in the final polls before the vote.
So those who said “Obama leads Clinton by X percent and he should win handily” simply assumed that those undecideds would break more or less along the lines of the decideds. They assumed wrong, and a good many reasons why this happened (race probably not being a serious factor, or not as serious as gender) can easily be imagined.
So trust the polls—that is, what the polls actually say–not their interpreters.
BTW, the exit polls that gave a slight edge to Obama were only off by a few points, and that’s not unprecedented. Remember President Kerry?
Also BTW, I don’t think this will carry far (see the post above on South Carolina). If Hillary’s crying moment helped, that has to be a blip. You don’t surmount doubts about your capacity to be President by tears. Endless replay of that moment in the café could, in fact, come back to haunt Hillary.
I have been watching Washington Week in Review for years and they are always surprised when the Polls get it wrong again.
The media went “show biz” on the job. They set up and displayed the circus that was N.H., blended it all with their own college days volunteer experiences, and lost track of reality. The fire spread, it gave them good ratings, they kept tossing on the gasoline. One even confessed while covering the last day in N.H. that “it is hard staying objective when covering this guy (Obama).” We knew it all along–just surprised she said it on air. There was everything in that circus but cotton candy vendors.
Condensed from my blog entry yesterday:
There is factor not much talked about. But this former political operative has no tact, so here goes.
People lie to polls. And there are occasions when they are particularly prone to lie to polls. If the media culture creates an atmosphere in which it seems people who don’t vote for a certain candidate are not “with it†or are even wrong-headed, some people will say they support said candidate even if they do not. (I’m convinced this was a factor behind inaccurate exit polls in the 2000 presidential election, by the way.) Peer pressure to conform or at least pretend to conform doesn’t end in high school.
Further, if the media-favored candidate is of African-American heritage, people who may not support said candidate are all the more prone to dissemble because they do not want to be perceived as racist. I know of one past Senate campaign so certain of this that they factored it into their evaluation of polls.
The media cheerleading of Obama was rather blatant. And, yes, Obama is of African-American heritage. So, yes, there was a lot of fibbing going on in those polls. Some people told pollsters one thing, while doing quite another when in the privacy of the polling booth. It’s not the first time that has happened, and it won’t be the last.
And not-so-by-the-way, the New Hampshire primary is indeed by secret ballot. The Iowa caucus is emphatically not. The Iowa caucus is a bit of a circus which compounds not dampens peer pressure.
Thanks, IRNS, I certainly like that analysis better than some I have read. I’m not convinced by the poll-lying analysis, which makes Democrats sound like either sexists or racists. Please. What it means is that polls in these open primary states where there are large numbers of undecided voters are of little use. And, if the exit polls cannot accurately predict close margins, then they are of little use either, except for after-analysis when their samples are adjusted to match the actual vote count.
All voters: vote your real opinion, and don’t pay any attention to pre- or exit-polls until you have cast your own vote. This will be important in the Super Tuesday primaries and in the general election.
I’m also, by the way, going to reject the thesis that Republicans are either sexist or racist when they vote against either Clinton or Obama in the general election.
And IF — IF — large numbers of Democratic and Independent women voted for Clinton *only* because they viewed her as a victim after her emotional moment, they ought to be ashamed of themselves.
Perhaps it all had rather more to do with a quirk of New Hampshire voting laws — it is possible to register as a voter on voting day itself [i]AND[/i] residence in the state is not necessary. Non-residents who declare their [i]intention[/i] to establish residence can be registered to vote.
The sample polls predicted the Republican side rather well, and Republican turnout was reasonable. Only on the Democrat side were the polls wrong. Only the Democrats had such a big turnout that they ran out of ballots in many places.
Ya think it mighta had something to do with the thousands of out-of-state cars that appeared on primary day? Nah. They musta been there as tourists.
It was vote fraud and vote suppression.
That’s the reason for the discrepancy between the polls and the election results. Obama was ahead of Clinton by 10 to 20 points in nearly every poll. But, after the primary, he was five or six points behind Clinton? On the Republican side, Ron Paul (who I’m supporting) was running 16% in the polls. After the primary, he only received eight percent of the votes? The “talkingheads†all want to blame the discrepancy on issues like racism, wrong exit polling and every other reason except the real reason. It was vote fraud!
I knew it was vote fraud the night of the primary and even before the primary. 75 to 80 % of the votes in NH were counted on Diebold voting machine. Diebold machines can easily be hacked and/or programmed for a certain vote outcome. You can google “Diebold machines hacked†or “Vote fraud electronic voting machines.†Thousands of mainstream and alternative news articles will pop up detailing the issue.
But, it gets better…..In the towns of Sutton and Greenville, they recorded that Ron Paul had 0 votes. People who voted for Ron Paul came forward inquiring why Ron Paul had a 0 vote tally when they voted for Ron Paul. The clerk that oversees the voting Sutton said it was an oversight and registered 31 votes for Ron Paul. The same for Greenville. Apparently, this has happened across the state.
On top of all of this, the guy who was the contractor for the Diehold voting machines is a convicted felon!!! He was convicted of drug running and distribution!!
So, there has been a big push in the alternative news community for a recount. Apparently, Dennis Kucinich has asked for a recount of the votes.
To find out more about the vote situation in NH, you can go to google news and type in vote recount and read the articles.
-Vincent
Maybe ” there are times that a cigar is just a cigar”. Hillary Clinton won, not because of the tears ( although it was a welocme sign that there might be a pulse under that waxy grin), nor because she’s a woman ( it must have driven the liberals nuts! Do I vote for the African American or the Woman? What’s a good liberal to do?). I think it is good old fashioned work. Hillary practices good old fashioned knock on the doors politics. While Edwards did his marathons and Obama his MLK immitation revivals, Hillary Clinton and her folks knocked on doors, over 90, 000 of them. And why did the pundits miss it? They were too busy with their lap tops, cell phone, and blackberries. The folks from the Belt way and NYC just did not get it: up in NH, a face to face cobnversation and handshake seals a deal.
I oppose electronic voting machines because of the possibility of programming error and lack of back-up. But I’d have to see evidence of hacking before I’d put the problem down to that. Also, as to the theory that out-of-staters came in large enough numbers to swing this is intriguing (what a terrible law!), but again, more than anecdotal evidence will be required. I hope someone is examining the voting rolls. Otherwise, IRNS’s theory that the undecided vote broke heavily for Clinton, and perhaps for McCain, seems more likely. Evidence, evidence, is what we need.
If the results of opinion polls were analyzed after the fact rather than before, I wonder if we would discover that many opinion polls were wrong more often than they were right.
Opinion polls are not as reliable as some might think because there is a human factor that makes them unreliable and that is the reluctance of people to participate in opinion polls. This is one area of our lives that we still have our right to privacy and more people are willing to say “no, I do not want to participate in your opinion survey†and “no, I do not give out information to a strangers who call me on the phone or send me mail asking personal questions.†This means that surveys are not of a full spectrum of the population but only of those who are willing to participate.
The only accurate opinion poll is the one performed in the voting booth.