When it came to Republicans, polls released before the New Hampshire Presidential Primary were pretty much on target.
They predicted John McCain would win.
When it came to predictions on Hillary Clinton’s performance, most polls were dead wrong.
New Hampshire Public Radio’s David Darman has more on why polls may have missed the impending Clinton victory.
Supporters at Clinton primary night headquarters in Manchester hold 'Fight On Hillary' signs that appear to anticipate a second place finish, January 8, 2008. Signs stating 'Ready to Lead' and 'Clinton Country' became more visible as the New York Senator maintained her lead. (Lauren Melcher, NHPR)
Hillary Clinton’s win in New Hampshire seemed to surprise a lot of people, including many pollsters.
But Dick Bennett of the American Research group in Manchester says that the media and the public may put too much faith in the predictive power of opinion surveys.
Polls don’t predict and if polls could predict the future, I wouldn’t be talking to you. Um, I’d be doing something else. No one can predict the future.
All polls do, says Bennett, is describe how groups of people are feeling at one particular point in time.
But Bennett’s pronouncements haven’t stopped his colleagues in the profession from wondering what went wrong.
Nancy Mathiowetz heads the American Association for Public Research in Kansas.
She says polling done right before the election couldn’t measure some of Clinton’s last minute support, because so many voters indicated they hadn’t yet made up their minds.
If you look for instance at the January 7th release from CBS news, they actually were citing in their press release that they, they’re finding that 28 percent of the Democratic voters said that they could still change their mind as late as the Sunday before the elections.
Other pollsters say a large number of undeclared voters also indicated they hadn’t yet settled on a candidate in the Democratic field.
David Moore is a former Gallup pollster who is now at the University of New Hampshire.
He says a UNH Survey Center Tracking Poll which ran through Sunday, January 6th didn’t catch how voters reacted to former President Bill Clinton’s attack on Barack Obama.
Moore says news reports on Sunday and Monday showed the former president saying Obama had said a few years ago he didn’t know how he would have voted on the Iraq war if he’d been in the Senate.
The reason I think that was so effective at least in my view, is that there was a series of attack on obama’s fundamental candidacy. And not only that, it went unanswered. Barack Obama never came back and was never able to rebut the specific charges that Clinton made.
Senator Obama did later respond to the attack, saying the former President had misquoted him, by leaving out part of what Obama had said.
Still, some pollsters say a last minute attack may well have been effective in driving last minute support to Hillary Clinton on Tuesday.
But those same pollsters say there is still much to learn on the surveying that went on before the primary.
They say it could be a few years before they can really tell the public why they were so wrong.
Do we have any idea how many "independents" voted for Clinton? It would not surprise me if republican operatives tried to make her win as she is more beatable than Obama. The ratio of republicans who would vote for Obama vs Clinton was hugely in favor of Obama!
A similar thing was done in CT with Lieberman election
Just my 2 cents
I just listened to your piece on the polling errors in New Hampshire. I have a different view on the errors. You have to live in NH before a primary to appreciate how many calls from pollisters are received along with the calls for the campaigns. I averaged 6-10 calls the weekend before the election. In discussing the situation with others, I came to realize how many of us hung up, pretended to be someone else or mislead the pollister in order to keep our sanity! To increase the number of calls and the length of questioning would no doubt increase the error rate. So please, leave the obsessive need for prediction and just let us independent New Hampshirites vet the candidates and vote.
There is a real possibility that this election, like the 2000 Florida and 2004 Ohio Presidential elections, was determined not by the voters of the state in which the vote took place, but, rather, by pre-programmed voting machines and ringers bussed in from out of state.
Why is the establishment media, which should be working to protect our democracy, so skittish about entertaining the entirely logical and eminently plausible notion that Clinton's miracle in New Hampshire was nothing more than common or garden variety graft -- election fraud?