How Many Polls Does It Take to Screw Up an Election?

"We've proven the experts wrong—again, and again, and again—and we're not stopping now!" said Mayor Michael Bloomberg in his victory speech last night. On at least one score, he was definitely correct: For the second election in a row, the most lavishly funded campaign in city history has totally befuddled the city's pollsters. Eighty-million dollars or more can win a mayoral race, apparently, but it can't beat the spread.

At campaign parties, in the on-air commentary as returns rolled in and in this morning's papers there were ample references to the strategic mistakes and personal flaws that allegedly hamstrung Bill Thompson's run. But what about those polls, which were way off, drove the media storyline and may have affected the election's outcome?

Four years ago, the two main city polls—Marist and Quinnipiac—had Bloomberg besting Freddy Ferrer by 34 or 38 points, respectively, in their final assessment of the race. Ferrer lost big, but not that big: The Election Day margin was 19 points. This year, the final poll by Marist had a 15-point gap and Quinnipiac's last survey had the gap at 12 percent. In the end, the mayor won by a mere 5 points. Polls showed a tightening of the race in the final days, which was apparently accurate, but were way off on how close it was.

The polls showing a landslide almost certainly tarnished turnout. Voters have more reason to vote when an election is close because their vote is more likely to matter, and New York City voters were repeatedly informed this fall that their vote wouldn't matter much. Of course, it's possible that both Bloomberg voters and Thompson backers stayed home because of the big gap in the polls, but the campaigns certainly treated the wide polling margins very differently.

In the closing days, Thompson kept saying that the race was "between three and seven points." His campaign publicized an internal poll that showed a gap of 3 percent among likely voters.

Bloomberg's communications director, Howard Wolfson, said Monday that "he doesn't spend a lot of time worrying about" polls. Yet last week, the mayor's camp released a goofy statement featuring musical notes and the memorable subject line "One of these polls is not like the others…one of these polls just doesn't belong." It contrasted Thompson's internal numbers to Marist and Quinnipiac. "Bill Thompson's poll gives new meaning to the term margin of error," Wolfson said in the statement. "Public polls show that the Mayor enjoys consistent double digits leads as we head into the final weekend before the election."

So Thompson's campaign was hyping a reason for voters to come out, while Bloomberg's was giving them an excuse to stay home. Not surprisingly, the better-funded campaign with the easier-to-obey message won out: Only 25 percent of registered voters cast ballots in the mayor's race yesterday, down from 29 percent in 2005 and 38 percent in 2001. It turns out that Thompson's poll was a much better predictor of Tuesday's vote.

The decreasing turnout during Bloomberg's political life means that a dwindling share of New Yorkers are actually voting for him: Given the low turnout and close race, about 13 percent of registered voters actually voted for the Bloomberg yesterday. In 1997, Ruth Messenger got the support of 16 percent of registered voters in her failed bid to unseat Rudy Giuliani.

Lee Miringoff, the director of the Marist Poll, defended the pre-election surveys in a blog posting today. "It is not unusual in contests between a well-known incumbent (Bloomberg) and a relatively unknown challenger (Thompson) that the incumbent ends up getting pretty much the same number he was attracting in pre-election polls. Undecided voters tend to find the challenger or not vote at all, having already rejected the incumbent," he wrote. "In the closing weeks of the campaign, all public pre-election polls had Bloomberg in the low 50s, regardless of the margin over Thompson."

But Marist's polls didn't reflect this caution. The survey released last Friday declared that "The race for New York City mayor is in the homestretch, and if today were Election Day, Mayor Michael Bloomberg would handily win a third term."

Besides, what the polls said wasn't as important as what everyone else said about them. The New York Post, for one, ran the October 30 headline "Mike a Near Lock in Poll" and one a week earlier that said simply "Breaking Away" and described a widening gap between the mayor and the comptroller. In the Daily News on election eve, Mike Lupica wrote that "barring some kind of miracle Tuesday, [Bloomberg] is supposed to beat Bill Thompson going away." Neither tabloid had the election on its cover during the final days, preferring instead to remind people the Yankees (another multimillion-dollar enterprise with no right to lose) are in the World Series.

The polls may even have helped convince President Obama to confine his comments on the New York contest to a begrudging minimum, kept other Democrats on the sidelines and discouraged donors from giving Thompson ammunition for the stretch run.

That's pure speculation, of course. And even if true, blaming the polls is unlikely to spare those wayward Democrats from the ire of Thompson partisans, who last night heaped almost as much scorn on their prodigal party brethren as on the billionaire mayor.

"We are Democrats, and those who betrayed us, we have to make sure that we don't forget," said State Sen. Bill Perkins. "Democrats, don't forget. Because when we forget, we get forgotten. We have to pay back."

Gov. David Paterson was only slightly gentler. "The fact is there are too many Democrats who stayed home today because they listened to the polls. They listened to people who represented everybody's interest but their own," he said. Bronx Borough President Ruben Diaz, Jr. said Thompson loyalists should be proud that they weren't "sell-outs."

Whatever effect the polls might have had on Thompson's chances at winning, the pre-election surveys only sharpened the righteous, sometimes angry narrative that Thompson's supporters wove last night. As was the case four years ago when Ferrer's performance defied the polls, Thompson supporters used the contrast between what the punditry expected and what the voters did to turn their man into a winner in defeat and Bloomberg into a vanquished victor. "I don't see a mandate emerging," Public Advocate-elect Bill de Blasio said. "Democracy has won a mandate tonight."

Everybody loves an underdog. Over at Bloomberg's super-slick celebration, after the mayor's victory speech, the sound system blasted Journey's classic anthem of battle against the odds, "Don't Stop Believin'." As if the polls had given the mayor's side any reason to doubt.

[Note: A previous version of this posting incorrectly stated that Mayor Bloomberg received 678,444 votes in 2005. This neglected his votes on the Independence Party line, which brought his total to 753,000. Like LaGuardia, Wagner and Koch, Bloomberg received more votes in his second race than in his first, and less in his third race than in his second. But unlike those earlier mayors, Bloomberg got fewer votes in his final run than he had the first time around.]