Clinton Surprise Brings Pre-Primary Polls Under Scrutiny

So much has happened in the past week it boggles the mind. Just a week ago, Iowa caucused and gave Senator Obama a surprising and definite victory to bring to New Hampshire. Senators Edwards and Clinton were retooling messages and machinery, and the press corps was flying to New Hampshire. Five days later, New Hampshire voters would weigh in.

Polls taken in New Hampshire stopped three days later, two days before the primary. Those polls were taken on the final weekend, which also turned out to be a good ski and tv sports weekend. The polls and, much more importantly, the primary, occurred before New Hampshire college and graduate students returned from break.

With Michigan holding to its January 15th date, New Hampshire's traditional eight day window after Iowa was compressed to five days. Secretary of State Bill Gardner picked the date as soon as Michigan resolved its date, just before Thanksgiving, and well after most students, faculty and staff of Dartmouth and UNH and numerous other colleges had made travel plans. Many students and faculty returned, but many did not. Those populous and progressive towns were strongholds for Obama.

Amidst the final polling, candidate events and media coverage were back to back ABC debates at Saint Anselm, the fierce fight between candidates, Bill Clinton on the trail, reports of a shakeup at the Clinton campaign, tears, attacks, likeability versus "you're likeable enough", the FOX News debate, and massive get-out-the-vote efforts. Perhaps the last thing the press could see is that its own conventional wisdom pre-Iowa would be the big surprise out of New Hampshire--a win for the candidate formerly known as inevitable.

Let's look at the numbers. The CNN/WMUR/UNH poll conducted 1/5-6 had a margin of error +/- 6% for the Democrats and showed Senator Obama with 39%, and Senator Clinton at 29%. The poll was widely used on Monday, January 7th, and verified by the results of other polls with similar findings. The key to understanding what the poll meant lay in something that went largely unreported: only 55% of respondents had "definitely decided" on their candidate. 45% of those polled were leaning in a direction or still deciding.

A FOX News/Opinion Dynamics poll conducted 1/4-6 with a +/- 4% margin had Obama at 32 and Clinton at 28. A Marist College poll conducted 1/5-6 with margins of error +/- 4% and +/- 5%, showed Obama at 34% and Clinton at 28%. When asked how strongly voters supported their candidate, 23% of Obama's supporters indicated they might change their vote, while only 12% of Clinton supporters indicated soft support. A 7NEWSBoston/Suffolk University poll conducted 1/5-6 with a margin of +/- 4% was the most accurate, showing Obama at 35% and Clinton at 34%. An American Research Group poll conducted 1/5-6 had a margin of error of +/- 4% and showed Senator Clinton 11 points behind Senator Obama. A Concord Monitor/Research 2000 poll conducted 1/4-5 had a margin of error for each +/- 5% and showed Obama and Clinton in a dead heat, 34 to 33. A Strategic Vision poll conducted 1/4-6 with a margin of error of +/- 4% had Obama 9 points ahead of Clinton. A USA Today/Gallup poll conducted 1/4-6 had a margin of error of +/- 4% and showed Obama 13 points ahead of Clinton.

So what does it all mean? It means New Hampshire voters did what they normally do: decide late, change their minds often, and don't talk much about who they are going to vote for. In an information age when most people use caller ID on their regular phones to screen telemarketers and their cellphones to actually talk, we may be past the point where voters can accurately be polled by telephone. That said, there was evidence that some polls with the highest leads for Obama also reported that nearly half of likely democratic voters were still deciding. And some smaller polls like those done for the Concord Monitor and Boston's 7NEWS weren't far off.

Exit polls aren't too scientific in a place where citizens keep their votes private. They did provide some good data points though. When asked when they made their decision, 39% of Clinton voters and 36% of Obama voters said "today."

Comments

[laocoon]

Given the drift this article and events show I suspect I will wake up some day to the following headline:
WIDESPREAD COMPLAINTS BY VOTERS IN ALL COUNTIES OF ELECTION FRAUD AND IRREGULARITIES CALLS INTO QUESTION THE HONESTY OF THE ELECTORATE.

[RTIII]

What a farse of an article!

Note that the meat of the article was about polling BEFORE the election. They, we suddenly get this switch near the end, "Exit polls aren't too scientific."

Exit polls are VERY scientific. We, the United States have on several occasions had elections overturned in other countries because the exit polls showed differently from the actual results by beyond a certain margin of error.

People tell you what they HAVE DONE much more accurately than they tell you what they WILL DO. HUGE difference in accuracy. Exit polls are to be believed. That the author posted this sentence shows either how uninformed they are or how deceitful they are. Take your pick.

Either way, if there was funny business to breathe new life into Clinton's candidacy, it would surely be because they - the Main Stream Media and the Fascist state they serve - wants the horse race to be Hillary/Obama and not Edwards/Obama because that would ensure Edwards message got main-stream attention and they could not easily avoid it!

This falls under the heading of, "Duh!"

[speakyourmind]

People, please: Polls have always been completely meaningless. Why is there so much fuss being made about this non-issue?

[theratt]

I'm quite tired of pundits poo-pooing exit poling. It is THE way to determine if the results in an election are viable. That's how it's done in every other country. Ask Jimmy Carter. Remember the Ukrainian election of 2004 where western countries including the US were calling fraud because of the discrepancy between exit polls and the final tally? Oh, the irony. There were counties in New Hampshire that reported zero votes for Ron Paul, for example, but there were people who specifically said they voted for him. Doesn't that fact alone call into question the official count? We're not talking about a 'bump' for Clinton, but a statistically unexplainable discrepancy in the polling. Interesting how the discrepancy does not occur in counties using PAPER BALLOTS. There need to be a recount and we MUST STOP USING ELECTRONIC VOTING! Oh America, welcome to friendly fascism.

[nurturingtouch]

If you love Kucinich consistent progressive stances, or love hand counted ballots. Then please donate to him with a note attached that the donation is to help support a recount. If it is correct then great, if not then it will perhaps be a powerful part of dismantling this terrible system that we have fallen in to. Its a Hillary or Obama thing it is a democracy thing. If the vote was skewed than those machines can skewer Hillary in 08, presuming that she is the eventual nominee.

[ptarantino]

Amazing numbers here:

I posted this once & it was rejected.

Clinton Optical scan 91,717 52.95%
Obama Optical scan 81,495 47.05%

Clinton Hand-counted 20,889 47.05%
Obama Hand-counted 23,509 52.95%

[Overd0g]

I would also consider that the vote was tampered with. Pols of the same party would never accuse each other of such a thing, but I wouldn't be terribly surprised, given the desperate situation she was in. After all, the Clinton's seem to idolize Jack Kennedy.

[tomass]

The repugnuts want to run against Hillary. That's all you need to know.

[letsmakesense]

What if the polls were correct?
Then the vote was hacked.
Why doesn't the MSM spend even 10% of their effort looking at that posibility?
Same answer as to why we still do not know the truth about JFK murder, Oaklahoma bombing, Gulf of Tonkin, American Union or 911.
VESTED INTERESTS

PS. Recount will show that Dr. Paul beat Rudy again and that Obama beat Clinton.
Can the rulers live with that?

[mellene]

I think as some have said that the polling ended too soon and/or didn't correctly model which way all the undecideds would fall. Probably as Clinton had been so far ahead for the majority of time in NH, it could have been predictable that voters wouldn't just move due to the caucus outcomes in Iowa. Really Obama went up a bump from an earlier period and Edwards fell. Clinton had always been near that 39 mark.

[YcCoSQ]

poll numbers should not be used to decide who is in debates.

[Canukistani]

I think that the only reason that this is such a big deal is that the pundits NEED the polls. The polls don't really do any great service to voters. What the polls do is they give the pundits something to pontificate about while they're waiting for something real to happen. They help the pundits to show you how superior their knowledge is to anyone else's and this justifies their existence.

When the polls go so wrong it's a threat to the pundits. If polls can be shown to be unreliable it's a real challenge to the pundits' illusion of superiority and their sense that they help shape events. If the polls can't be trusted people can see that it's all nonsense. That's why there's all this hand-wringing and looking for something or someone to blame.

[fourjays30]

Okay I am going to finally post on this web site. I have been coming here for over a year just to read the articles and the comments. Finally after reading the comments about the vote in NH and all of the mis information that is being put out there I thought it was my duty to set the record strate. I live in NH and I voted in NH last Tuesday, How many of you that say the results were rigged can say the same? First of if you knew anything about the state you would realize how small it is and how many of it's communites vote with paper and pencil and not voting machines. There are only a few cites large enough that vote with machines. The results after the election was pretty similar in all of the towns accross the state. The boston globe had the results broken down by each town on each of the two races. Her percentage of the vote was about the same in the towns that used paper and pencil and those few that are large enough to have voting machines. The only places wher she did not carry the vote was understandbly college towns which were the hot beds of his supporters. Please if you do not know the facts and you did not see first hand what happen on election day do not spout comments about voter fraud because all that will do is when there happens to be real fraud no one will believe you.

[nellie]

It's not the pre-polling that's the problem. A large percentage of people polled before the election said they hadn't decided.

It's the EXIT polls being so different from the outcome that raise the red flag. Exit polls ask people what they actually did. It's factual information being collected.

I wish the bloggers on HuffPo would get this straight. Not one of the articles here seems to understand this simple concept.

[DietrichMoody]

What Clinton Surprise? The woman has been ahead in every poll for over a year, and after she lost Iowa people started to wake up and say, hey, maybe she isn't inevitable, hey the world didn't end after she LOST. That was a good feeling for a lot of Democrats that never felt that had a choice other than Clinton. And its ironic that all who cry "choice" "choice", cry "foul" when people exercise their choice other than Hillary Clinton.

Besides it was the Clinton Machine, via 1992 that elevated and took to a whole new level, this whole incessant dependency on THE ELECTION as a alleged War sport with, polls, pollsters, pundits, etc. thus scripting the " the perfect candidate with the perfect pedigree"
There are good candidates with bad "pedigrees", lets find them and let it not be a Clinton. Hey Eisenhower, Truman, and Ronald Reagan are great examples.