I'm still trying to figure out who had a more wretched Election Night 2002, the Democratic Party or America's pollsters. While Democrats lost control of the Senate, they will live to fight another Election Day. Pollsters, on the other hand, in losing what scraps of credibility they had, may -- with a little help from the public -- find their entire profession obsolete, gone the way of chimney sweeps, organ pumpers, and those guys who used to make buggy whips.
For years now, the accuracy of political polls has been -- in the parlance of the trade -- "trending downward." Last week it hit bottom. The Voter News Service admitted on Election Night that due to "technical difficulties" its exit polls weren't to be trusted, forcing the networks to rely on actual votes. And in race after race, preelection polls proved as reliable as the iceberg spotter on the Titanic.
In Georgia, pollsters had predicted Gov. Roy Barnes, the Democrat, would handily beat challenger Sonny Perdue: A Mason-Dixon poll had Barnes leading by 9 points, while one conducted by the Atlanta Constitution had him up by 11. Once the votes were counted, however, it was Perdue beating Barnes by 5 points -- a humiliating 16-point air ball for the pollsters.
They were just as prescient in Colorado, where an MSNBC/Zogby poll had Democratic challenger Tom Strickland trouncing incumbent Sen. Wayne Allard 53 percent to 44 percent. In reality, Allard strolled to a relatively stress-free 5-point win -- a 14-point blunder.
And in Illinois, another Zogby poll had the governor's race pitting Republican Jim Ryan against Democrat Rod Blagojevich as a statistical dead heat -- a finding that was, statistically, dead wrong. Blagojevich won, and Ryan and Zogby lost by 7 percentage points.
The pollsters' numbers were so off the mark that even they were forced to admit the obvious. "There was a lot of bad polling this year," acknowledged Mark Mellman, a Democratic pollster. "We blew it," said John Zogby.
As a rule, pollsters come equipped with more excuses than a married man with lipstick on his collar -- and this year was no different. And whom did they point the finger of blame at most frequently? Why you and me, naturally. It seems we just didn't turn out at the polls in exactly the configurations the pollsters thought we would -- what Mr. Zogby delicately referred to as "poor turnout models." In other words, the problems aren't polls, it's those damn voters who say one thing then do another. Like show up on Election Day.
In truth, the problem isn't with us, dear voters -- or even with you, dear nonvoters. The problem is with the pollsters' inability to account for an increasingly uncooperative public. Thanks to cell phones, answering machines, caller I.D., a surfeit of polls, and a growing distaste for telephonic intrusions into our homes, it's getting harder and harder for pollsters to find Americans willing to answer their questions. Twenty years ago, polling response rates were over 60 percent; now they are closer to 30 percent -- and in some cases even lower. It's pretty tough to get an accurate reading of the public's opinion when the most frequent response you receive is a "click" followed by a dial tone.
So here we are in the middle of a vicious vortex. Pollsters conduct their increasingly inaccurate polls; the media then report the results as if Moses has just brought them down from the mountaintop; and our politicians tailor their messages to suit phantom voters. All the players involved in this charade understand they are acting on the flimsiest of pretenses -- it's just that relying on polls is so much easier than actually reporting or leading.
Even President Bush, who charged into office trumpeting his disdain for polls -- don't they all? -- has proven to be a chronic poll watcher and poll taker. In fact, this schizophrenic stance has actually become something of an in-joke at the White House: Bush brags about not looking at polls and everyone laughs, knowing the president doesn't have to look because Karl Rove has already whispered the results in his ear.
But allowing polling data to become a substitute for thinking has become a very wobbly crutch indeed. Just ask the Democrats who, after consulting their pollsters' tea leaves, decided not to take on the president on tax cuts or on invading Iraq. They were forced to pay for their slavish devotion to the numbers with their political lives. Pollsters, on the other hand, are allowed to tiptoe away from the carnage their handiwork has wrought and still keep their jobs.
As long as you can sagely and entertainingly spin your numbers on the tube, there is no penalty for being wrong. As Norm Ornstein of the American Enterprise Institute wryly puts it, "It's the sin of broadcasting in the modern age. No matter how wrong you are, the punishment is you get your own show on cable television."
I think it's time to change that equation -- to attach some downside to the political prognosticators' game. Perhaps we should fine pollsters $100,000 for every percentage point they are off (and create a retirement fund for pollsters who agree to leave their discredited profession). Or attach a large letter "I" (for "Inaccurate") to the lapels of those who are wrong more than they are right. Or perhaps we can follow the lead of English soccer leagues, which regularly consign teams with losing records to second-tier divisions. And if all else fails, there is always the option of a little reverse Pavlovian training -- let's say, by attaching electrodes to pollsters' sensitive areas on Election Night and sending a charge through them anytime a poll-based prediction proves erroneous. It would give a whole new meaning to the term "political buzz."
If you, like me, are one of the many millions who hang up on callers wanting to know what kind of toothpaste I prefer, what TV shows I watch, or what candidates I'm going to vote for, you'll be proud to know that you are part of a rapidly expanding segment of the population known as the "margin of error." And if you're not, now is the perfect time to join us and make antidemocratic polling a thing of the past.
Shares