There’s an easy way to predict election results: follow the polls. The candidate with a statistically significant lead closest to the election will almost always win.
It’s such a simple formula, but one that eludes partisans of all stripes. So much so that there’s a second near-surefire way to predict electoral outcomes: the candidate who says “the only poll that matters is on election day” or whose supporters come up with excuses for why the polls are wrong in their particular case (Our voters prefer cellphones! MSM liberals/corporate media reactionaries are oversampling Democrats/Republicans!) will almost always lose.
There are exceptions, of course. But as someone who was trying to guess what would happen in politics at an age when normal boys were playing Little League, I can tell you I’ve seldom gone wrong following the polls. My correct predictions contrary to the polls have been much rarer.
So the news that Gallup is getting out of the primary polling business, and will perhaps sit out the general election, this cycle was a major upset in its own right. Perhaps it shouldn’t have been, given the venerable firm’s shift in focus to things other than the horse race.
In 2012, I followed the polls to a split decision. The state polling looked good for President Obama, but Gallup had Mitt Romney up by 1. Don’t be surprised by another popular vote/electoral vote mismatch, I said, this time in the Democrats’ favor.
Whoops.
Polling may still be the most accurate way to predict what’s going to happen in an election, but the industry has serious problems. Response rates are abysmal. Cell phones are indeed replacing landlines (according to one estimate, a landline-only poll in 2014 would have missed nearly 60 percent of the American public, three times as many as would have fallen through the cracks in 2008). It has always been difficult to figure out which respondents will actually vote.
Reputable pollsters have methodologies for getting around each of these problems, but some of them are expensive, raising the price of good polling, and none of them are foolproof. As a result, we have seen some polling struggles.
Some organizations underestimated the Republicans’ strength ahead of the 2014 midterm elections. Polling conducted abroad similarly underestimated support for the Conservatives in Great Britain and the coalition behind Israeli Prime Minister Benjamin Netanyahu.
Even before some of these problems started coming to a head, there were problems polling presidential primaries. “The average error for presidential primary polls since 2000 has been 7.7 percentage points — about twice as large as for presidential general elections,” observed polling guru Nate Silver. “The polls were especially bad in the 2012 Republican primaries, when they missed by an average of 8.7 percentage points.”
This wouldn’t matter much if only columnists’ political predictions were affected. But we’re increasingly using polls to shape the presidential race rather than simply measure it, especially in the primaries.
For Republican presidential candidates this year, making it into the primetime debates rather than the undercard events for lower-polling aspirants has been the difference between success or failure. Just ask John Kasich, whose campaign was materially helped by his inclusion in the first debate in Cleveland, or Rick Perry, who was fatally damaged by not making the cut. Even the possibility of a demotion helped end Scott Walker’s once-promising and well-funded (at the super PAC level, at least) candidacy.
Republicans needed some way of sorting a field of sixteen or seventeen candidates in a way that would allow for a reasonable exchange of ideas. But the low-polling contenders who complain that the networks’ standards seem a bit arbitrary, especially since they are based on national polls rather than surveys of the states that will actually award the delegates, have a point.
Polls that may be getting less accurate are being asked to do more to help winnow the field and influence the process. Even insofar as these polls are right, they are just snapshots in time. The results this early out are subject to change. Just ask Presidents Giuliani, Bachmann and Cain. In Donald Trump, we now have a candidate whose strong performance in the polls is a core part of his campaign stump speech.
We’ll have to see what it means for the all-important poll on election day.
