The great poll crash of 2008

“America in Search of Itself” is what the late Theodore White called a book of his about the presidential elections between 1956 and 1980. The title’s not a bad description of what happens every four years, of how we look at ourselves and candidates and try to figure out where we are, where we want to go and who might lead the way.

It’s sometimes an exciting exercise, but inexact and exasperating. Among those who make it less endearing are the pollsters telling us what we will learn soon enough anyway and members of the press who think it their bounden duty to dwell obsessively on strategies, tactics and just about anything except what the candidates are saying and its possible resemblance to any discernible reality.

In the Democratic primary in New Hampshire this past week, much of that was smashed. Happily, wonderfully, marvelously smashed.

Nine different polling organizations had Barack Obama winning, and winning big. One of them, Gallup, showed Obama galloping across the horizon with as much as a 13-point advantage. Hillary Clinton was supposedly washed up, finished, a mere wave of the hand from the distant, pre-primary past. The outcome, we all know, was different. Clinton won with 39 percent of the vote. Obama had 36 percent.

So how did the polls go wrong? The answer is that they probably didn’t. I myself read polls voraciously, and I read them in part because I know the best of them are highly trustworthy. I also know they have their limitations.

Polls are not about the future, which is what people like me often want to know. They are about the times during which the polls were taken, and even if that is just hours before you read the results, you are reading about the past in a world that happens to be dynamic, not static. Things can change. Even if there were no other possibility of error — and there always is in anything self-reported — this fact alone would mean that no poll can be taken as the final word of what will happen in the voting booth.

Next question: Do election polls do wrong? Well, there’s this: They are an investigation that affects what they investigate. Through a variety of ways, they can change the outcome of an election and may have this time around.

Stories tell us that Clinton advisers, believing the polls, saw their candidate in deep trouble. No doubt, she did, too, and her fear of ultimate failure may have led to the moment when she almost came to tears in response to a coffee shop question about how she was able to keep going. It wasn’t easy, she said, but she believed in trying to keep the country from falling backward.

“This is very personal for me … it is not just political,” she said in an injured, vulnerable, sincere voice. The thought is that, after seeing this moment on TV, thousands of women were moved to vote for her. I believe it. I am far from being a pro-Hillary Democrat, but I saw it and found myself sympathetic. For maybe the first time ever, I liked her.

Reporters, following not just the polls, but also the herd and their druthers, had pretty much crowned Obama not only the New Hampshire winner, but something close to sure-fire in capturing the Democratic nomination. Editors could avoid these embarrassments by instructing their troops to focus on factual stories related to policy, character and competence, with far less emphasis on strategies, tactics and predictions.

The New Hampshire experience will not change reporting overnight, but maybe it will make some reporters a little more humble, a little more inclined to self-examination. Nor will the experience make polls go away. They shouldn’t; in many ways, they serve us, especially in giving us a more complete picture of the electorate than we could otherwise get.

But if the failure of the polls in the Democratic primary causes people to suspect them a little more, to feel, for instance, that their vote might help make a difference even if the polls indicate otherwise, it will be for the good of all of us as we continue to search for ourselves in 2008.

(Jay Ambrose, formerly Washington director of editorial policy for Scripps Howard newspapers and the editor of dailies in El Paso, Texas, and Denver, is a columnist living in Colorado. He can be reached at SpeaktoJay(at)


  1. Cobaltkid


    Your conclusions are suspect since your highlighted source does not accurately reflect the method of vote counting. My own town, for example, is shown as hand counted when it is actually done by machine. How do you know that the results were fixed when your source contains inaccurate data?

    The polls were wrong for several reasons such as:

    – Many were conducted and structured very poorly to the point of being unprofessional and worthless.

    – There were probably too many people, especially new young voters who believed the early polls and didn’t bother to vote.

    – The pollsters didn’t examine the traditional democratic strongholds of Manchester, Nashua, Salem, Berlin, etc. where the Shaheen machine performed it’s usual well organized operation of getting out the traditional democratic base.

    – Since the whole polling process was jammed into a few short days, voters were getting 10 to 15 calls a day that lead to total exasperation. Some of these calls ranged from recorded polls conducted by enemies
    of current congressional office holders to horribly prepared callers who probably couldn’t have located New Hampshire on a map, let alone understand the local polical structure. Many people simply hung up or as one of our friends said, tell them what they want to hear so that you can get off the phone.

    The whole process makes me a proponent of banning ALL pre-election polling.

  2. Cobaltkid


    New Hampshire Secretary of State Bill Gardner just announced that there will be an official recount as soon as the candidates requesting same send funds to cover the cost, which is being estimated.

  3. KayInMaine

    Cobaltkid, Dennis Kucinich is requesting a recount. Even thought he only got 2% of the votes, he does have a right to have a recount, will be sending the money to have it done, and is doing this because he believes Obama was gypped.

    We’ll have to wait for the results. The whole thing did feel fishy and if computerized voter fraud is happening this early in the game, then something has to be done NOW.

    Just like the 2006 mid-term election, attorneys from both parties need to be present to watch the counting.

  4. keith

    Folks, as I’ve stated in the previous thread, all this “Diebold conspiracy” bunkum is just that…. bunkum!

    Rather than constantly repeating this dogma, I suggest those who continue to cling to such baseless blather might want to first do a little more statistical digging. You might learn that the apparent “irregularities” that emerged in New Hampshire last Tuesday don’t AT ALL prove what you may think they’re proving.

    In fact, what the statistics DO show is a repeat of voting patterns that pre-date the use of vote counting machines and that have been very much a part of elections in our State for decades. Factors such as education, income, age and where people choose to live are all factors that influence voters’ choices here.

    For example, in 2008, 2004 and in 2000, Democratic voting from those (usually larger) towns and cities (which just so happen to also use ballot-counting machines) were skewed toward Democratic primary winners Clinton, John Kerry and Al Gore. Conversely, overall second-place finishers Obama, Howard Dean and Bill Bradley won contests in towns where ballots are hand-counted.

    That is, the EXACT SAME VOTING PATTERNS that have been observed in years past in New Hampshire again emerged in the vote last Tuesday. But, what is probably even more telling is that ALL of these past results remain consistent with a number of independent exit polls that were conducted by a number of organizations last Tuesday as well.

    It is also not surprising we’d see such patterns emerge when we compare them with the types of vote counting equipment used. That’s because such equipment is NOT randomly assigned throughout our State. That is, there are good and valid reasons why certain towns and cities use paper ballots and certain (mostly larger) cities use machines…. mostly having to do with the volume of voters in each precinct.

    Manchester, for example, is New Hampshire’s most populous city. It is largely a working class city and it uses vote-counting machines at its 12 polling stations. Mrs. Clinton won handily there Tuesday, just as previous winners Kerry and Gore did.

    On the other hand, the small northern towns of Franconia, Sugar Hill and Bethlehem, which hand-count paper ballots, all went to Obama as they did for Dean in 2004 and Bradley in 2000.

    What’s more, Clinton, Kerry and Gore ALL seem to have attracted voters that fit a similar profile in New Hampshire. These voters have tended to be older, less likely to be college-educated and (on average) lower wage earners.

    These are what I call the old time, “straight ticket” Democrats. They live mostly in the larger cities in our State and have consistently been voting for “next in line”, labor-union-endorsed Democratic candidates (like Clinton, Gore, and Kerry) for decades.

    On the other hand, those persons who voted for the so-called “upstarts”… Dean, Obama and Bradley…were most likely to be younger, have college degrees or postgraduate education, and be in a higher income bracket. And, as one would expect, these people tend to live outside the larger cities.

    Also, in terms of issues, the divide this time was pretty much based on a candidate’s experience versus change. That’s not unlike the split between Kerry and Dean was four years ago (and between Gore and Bradley eight years ago).

    It is also important to remember that ALL ballots in New Hampshire elections are cast on paper. It’s only some of the COUNTING that is done by machines. Roughly 58 percent, or 175 of New Hampshire’s 301 precincts, count their paper ballots using AccuVote optical scan machines. These are the only voting machines that have been approved by the State. Touch-screen voting machines are NOT used in our State…ANYWHERE.

    So, once again, the only “conspiracy” (if it can be called that) in the voting on Tuesday was simply us New Hampshire voters being our normal, independent and crusty selves…voting our conscience while taking sheer delight in confounding all the pollsters, pundits and talking heads…not to mention all you voting conspiracy theorists!

    The patterns in our voting have traditionally had FAR more to do with who we are and where we live rather than some underhanded scheme to tamper with our State’s vote counting machines.

    Or, to put it another way, we New Hampshire hicks may appear a bit backward at times, but we certainly aren’t stupid.

  5. Cobaltkid

    Keith – Great Points ……….

    Your comments do a much better job than mine of explaining NH demographics and voting patterns. Maybe I am being naive, but I in no way believe the accusations that are being hurled at us. Who the H… is this guy from Detroit who flatly stated on national TV that he thinks the election was rigged! And no, we don’t need legions of lawyers, a la Florida, coming in to tell us how to count votes.

    In talking to our senior Supervisor of the Checklist, I learned that Concord has only approved two voting machines. I believe our town uses the Diebold, which is a stand alone machine.

    It will be interesting to see the results of a recount. I will go on record as saying that the results will be very close to the original tally.