A common activist slogan during my college years was "What if they gave a war and nobody came?" A bit of tortured cleverness, sure. But it has made me wonder, with some wistfulness, what it would be like if we had an election and there were no polls.
Opinion polls used to be secondary factors in a campaign narrative at best. They provided an occasional snapshot about where the public was on candidates and issues, and were largely there to aid the campaigns themselves. Candidates and their handlers could see where they were strongest (and weakest) geographically, and what issues were resonating with voters. But the continuous onslaught of polls—some more reliable than others—has turned the numbers game into the dominant force in the campaign narrative. And worse, the numbers are driving the narrative now more than they are illustrating it.
A few weeks ago, the numbers were bad for Mitt Romney. President Barack Obama had hosted an unexpectedly fiery and upbeat convention, and for a while, his numbers looked good—if a little suspiciously so. In Ohio, a state that has been razor-close in recent elections, one poll had Obama up by double digits. That number may well have been accurate—but merely for what it was, and for what all polls are: a snapshot of one place, at one fairly limited time. Obama had been using Romney's opposition to the auto industry bailout to his advantage, and it appeared to be resonating. But 10 points? That's a big lead in a state that other indicators show is more evenly divided.
But since Obama's numbers were good (then), Romney's campaign missteps—and there were, indeed, missteps—were presented by the Democrats, some irritated Republicans, and the media as a sign of a campaign in Titanic-like descent. Romney was done, dead. There was almost no real reason to have an actual campaign. We had the polls.
Except that oops, Romney came back—and Obama started to fall. This was also somewhat event-driven, since the narrowing of the gap followed an abysmal debate performance by the president. Yet polls also showed that most people's opinions of Obama didn't change much after the debate (even among those who acknowledged how lackluster Obama's debate showing was). If the polls swings show anything, it's that a strong performance in an event (convention, debate) will make people feel better about a candidate about whom they had regarded with some skepticism.
That's not what we hear on TV or read in much of the print press (excepting the great Nate Silver, author of the 538 blog in the New York Times and one of the few voices of sanity in the polling analysis business). The narrative now—driven almost entirely by the polls—is that Obama is tanking, in grave danger of losing states like Pennsylvania that only two weeks ago were considered fairly reliably in the president's column. But not to worry, Obama supporters—the TV pundits also tell us that the coming debate could change everything. Again! Until the next debate, which could also change everything.
What would the campaign coverage be like if there was no polling? I've experienced this, on a smaller scale, covering elections abroad. In Albania in the 1990s, an armed uprising by disaffected citizens (who raided the armories for weapons) led to what would have been inconceivable a year before—a serious challenge to then-president Sali Berisha. Legions of Albanians had lost all their savings through a bogus pyramid schemes, and many blamed Berisha, whom they felt had given tacit approval of the schemes. Would Berisha's political machinery keep him in power? Would voters reject the leader so many of them blamed for their financial losses? How to assess political behavior in a country that still had family "blood feuds"?
We didn't know, since there was no polling. The narrative of the race was not about numbers, not about strategy, not about issues messaging. It was about the issues and about the people. And while many of us sensed Berisha would be punished (and he did lose, though he returned as Prime Minister in 2005), we simply didn't know what would happen until the votes were counted.
Polls can be illustrative and useful. But they should take a back seat to reporting on actual issues. No matter how enticing a one-point shift in polls might be to candidates and the press, there is really only one polls that matters. And that is the one on Election Day.