Ten Miles Square


November 14, 2012 10:50 PM The Great Campaign Polling Conspiracy

By Michael Kinsley

It’s a small matter, I know, compared with the historic issues now obsessing the commentariat, such as the fiscal cliff and how many mistresses and admirers former Army General David Petraeus could keep in the air simultaneously.

But before we say goodbye to the Campaign 2012, I would just like to point out that the entire drama of a close election, as played out in the news media on Election Day and evening, is basically fake. Like broadcasters (including a young Ronald Reagan) presenting baseball games in the early days of radio, the television networks know who’s going to win the game and more or less how it’s going to play out, inning-by-innning.

They know this primarily because of research conducted by the National Election Exit Poll on Election Day. And yet, in a perverse exercise of high-mindedness, the major news organizations have all agreed not to report the results of exit polls until after the polls have closed in a particular state.

It has evolved into a semireligious ritual. At 11 a.m. on Election Day, representatives of ABC, CBS, CNN, Fox, NBC and the Associated Press entered a “quarantine room” with no telephones or Internet access. There, they sat and analyzed the exit polls until 5 p.m., when they released what they had to their employers, who got the data directly for the rest of the evening.

No Predictions

Exit-poll data is supposed to be used for demographic insights only — not to predict the result. You can say, “Republicans are doing well tonight among upper-middle-class white men aged 35 to 45, wearing a red sweater vest and answering to the name of ’Champ.’” But you can’t say, “Chances are better than even that Obama’s got it in the bag.”

You can learn a lot from tiny samplings, comparing them with past results. By 6 on election night, CNN undoubtedly knew that President Barack Obama was going to win re-election. And they pretty much knew the Electoral College count. Or at least they knew it reliably enough to want to deny this information to their viewers.

Thus there was this stilted dialogue, airing sometime between 6 and 7 p.m. on CNN, between John King and Wolf Blitzer about Vigo County, Indiana:

“One little ad-lib here, if I can,” said King. “We’re starting to get results in Kentucky and in Indiana. Tiny results, 2 percent of the vote. I want to show you a little place in Indiana. Vigo County, 1.7 percent of the population. … Only twice — only twice since 1888 has Vigo County been wrong in picking a president. Why? Good question. But since the 1950s, this county has been right. It’s filling in blue at the moment. Look at that. That’s only 17 percent of the vote. We’ll see how it goes tonight, but you watch it blue now. If it’s blue at the end of the night, we’ll see if Vigo County’s streak continues.”

King seemed to be saying that if Vigo County stayed blue (that is, voted Democratic), it would continue its streak of picking the winner. That would seem to imply that Democrats were going to carry the day. King arguably saved himself at the end with a “we’ll see how it goes tonight,” but he sure sounded like someone assuming a Democratic victory at a time when he and everyone else were telling viewers that the race was too close to call. (For those of you scoring at home: Vigo County went for Obama by 339 votes.)

Blitzer then said thoughtfully: “Could be a bellwether, as they say; could be an indication of what’s going on. We’re going to watch all these states, all these counties, all these polling precincts very closely.” Then he tossed to Anderson Cooper, who added, “Who knew?”

The answer is that all three of them knew, or someone in the studio with them knew. But they were forbidden to say. When I worked at CNN, I was even forbidden to say that I was forbidden to say.

No Reporting

This is not merely an American insanity. In some European countries, reporting the results of exit polls (or sometimes of polls taken close to the election) is actually a criminal offense. The reason is that reporting the result while the polls are still open somehow devalues the votes of people who haven’t yet voted. This might discourage turnout, and even change the result.

Is this a valid concern? No. Now children, listen closely: Your vote is just as valuable — or, if you prefer, just as worthless — no matter when you exercise your franchise. Get over it. No national election (not even the 2000 presidential election) is ever decided by one vote. If it ever were, every voter at all times of day would be equally implicated. Exit polls can’t predict the outcome of a contest that close anyway.

If it bothers you that the result has been decided before you cast your vote, that unfortunately will still be true whether the exit polls — and the conclusions experts draw from them — are made public or not. Yes, the polls and experts can get it wrong. But the concern here is that they usually get it right. How can it devalue your vote to give you information you wouldn’t otherwise have? What unfair advantage does an early- morning voter (or someone who voted weeks ago, absentee) get from his or her lack of information?

Yes, voting is a good thing and should be encouraged. But people shouldn’t be tricked into voting, which is what this artificial suppression of information amounts to. And yes, it’s possible that some people — rationally or otherwise — will decide not to vote if the winner has already been announced. But there is no reason to think that one candidate’s supporters are more likely than another’s to drop out, so that this could change the result.

It’s easy to see why the TV networks don’t mind putting on a play if the suspense keeps people watching past 6:30 p.m. Especially when they get civic brownie points for doing so. And why is this so important? Maybe it’s not so very important — a writer needs some hobbyhorses, and this is one of mine. It amazes me that, with the encouragement of the government, not to mention an endless string of foundations and commissions and pompous individuals, some of the biggest players in the media business conspire to present a view of the world that they know to be false.

It’s as if the government staged the whole walk-on-the-Moon thing in a warehouse somewhere, or as if Obama was born in Kenya. Except this one is for real.

Back to Home page

tags ,


  • Crissa on November 15, 2012 5:50 AM:

    I'm old enough - although I was a child at the time - to remember all the exit polls being released early in November, 1980.

    Nearly no one went to the polls in Oregon, where I was living.

    It was a very disappointing night.

  • Lucia on November 17, 2012 10:07 AM:

    The concern as I understand it is that if the presidential race is called before 11 pm EDT, no one on the West Coast will bother to vote, and while that doesn't matter to the presidential race it can to down-ballot races.

  • Sam Wang on November 17, 2012 10:56 AM:

    Crissa's recollection is incorrect. As far as I am aware, this is an old canard.

    Based on data at Michael P. McDonald's site http://elections.gmu.edu/voter_turnout.htm
    the turnout in 1980 for highest-office voting (the Presidency) was an average of about 56% nationally, and 63% in Oregon. Whatever the faults of Election Night news reporting, I question whether significant suppression of West Coast voting is among them.

  • Roger Keeling on November 17, 2012 11:51 AM:

    Lord, I don't want to argue with Sam Wang about election turn-outs! But like Crissa, I too remember the 1980 election (I was in Santa Barbara, CA. then). I don't agree with her that it was the exit polls that caused us much grief; rather, it was President Carter's extremely swift concession -- hours before our polls closed -- that distressed us so much.

    No, I can't recall many specific local or state races. But I do remember that after the concession speech, a lot of close races across the state went to the Republicans ... more than we'd expected. In Santa Barbara we had at least one very close county race we'd been optimistic about go against us; right or wrong, we widely blamed Carter's early concession as the cause.

    This wasn't a matter of whether our turnout equaled or exceeded the national average, but only whether it was suppressed at ALL over what it would have been otherwise. If it was, then our distress and anger over Carter's early concession (and exit polls, too, for that matter) was in fact legitimate.

  • Steve K. Johnson on November 17, 2012 7:01 PM:

    I remember the Carter/Reagan election. People wanted to vote for Reagan that year, just as people made a special to cast a vote for Obama in '08. The issue is that many people are most aware of the national race and may have only a minimal opinion about the races farther down on the ballot. With an announcement of the results of the Presidential election, it is the state and local candidates that can be negatively impacted.

  • brian t. raven on November 17, 2012 9:41 PM:

    Can we get some of the Washington Monthly Poly Sci gurus to weigh in on this? It's an important issue that shouldn't just be left to speculation.

  • Roger Keeling on November 18, 2012 3:03 PM:

    Hey, Brian: Sam Wang IS a guru on this topic. He's one of the biggest of the heavy-hitters on polls and their analysis, with an impeccable record. He's found over at the Princeton Election Consortium. He was widely cited and followed prior to this most recent election.

    Which is why I said I really did NOT want to try arguing with him, since he probably knows something on the order of a 1000 times more than I do about the topic generally. My quibble -- echoing Crissa up above -- was that our perception in 1980 was that we lost some or many down-ballot races on the West Coast due to already-discouraged voters just chucking the whole exercise when they learned hours before the polls closed that Carter had conceded. Maybe we were wrong, but we certainly believed it at the time, and had a lot of razor-thin races we'd lost as evidence. Early announcements by the news networks of "the winner" based on exit polls from the East Coast might -- some of us feel -- contribute even more to this sort of thing.

  • Aussie on November 18, 2012 4:07 PM:

    This column reminded me of a quibble that I had with your election night coverage – only rarely did CNN, or Fox when I switched over to watch them squirm, tell their viewers how the incoming results compared to the 2008 results in specific geographic areas.

    It was utterly useless for someone with a very limited understanding of American political geography (and that would include most Americans, not just overseas viewers) to know that Obama was 20 points up in Ohio as early precincts reported. But it would have been very informative to know that Obama was, say, 4 points down on his 2008 results in those precincts that had reported when Romney needed him to be 6 points down across the state (don’t quote me on the numbers). Same goes for Virginia where Romney was up 20 points early. I understand that the early reporting areas “swung” harder there than the later reporting areas, so it wouldn’t always be a 100% accurate forecast of the result early on, but it would have been better than just showing a 20 point margin.

    The networks seem to have access to precinct level data, so I don’t see why they couldn’t produce an average swing quite quickly, to compare against the required statewide swing. The best Australian analysts do it on our election nights and it’s very helpful. I wonder whether the fear of calling the race too early for the West Coast stops the American networks, or whether it’s some other issue, like keeping everyone watching?

  • N.Wells on November 18, 2012 5:28 PM:

    In this instance, I don't care what the research says (but I'd be surprised if it didn't show effects on down-ballot races) - it is simply a horrible idea to be announcing outcomes before people have finished voting. I don't even like eastern states being called while western states are still voting.

    That being said, the solution is that we need to fix our voting schedule. It's not good that polls are closing at as early as 7:30 in some states, nor that we vote on a work day. We need a National Voting Day, a business holiday, perhaps a Saturday, complete with civic celebrations of democracy, as well as the business of voting. It would also not be impossibly difficult to hold elections over a 24 hour period that begins and ends at the same moment nationwide (even with night hours off), which would allow networks to call the election quickly without discouraging any voters.

  • grandpa john on November 18, 2012 7:19 PM:

    Who bothers with exit polls when you can be getting real information.Hell, for those of us who follow Sam Wang on his Princeton blog and NYT's 538 blog , we literally knew months ahead who was going to win,and that it was not going to be a horse race. I certainly don't depend on MSM whores like King or Blitzer for any worthwhile information.

    Also what happened to the promised changes of the stupidest thing here, captcha. Now that the election is ever I see no reason to continue to torture myself with this blithering piece of nonsense, and If I am not going to comment, they why bother to even come here at all.