We're at the stage of the election where obsession with the latest polls reaches its peak. Yet it is also the stage when polls can stop informing about the electorate and start warping the electorate. We can't pretend polls don't exist, but we should handle with care.
The most important thing to know about election season polls is that the ones that generate headlines use "likely voters." But there is no single set of criteria that determines who is a likely voter. Pollsters use different approaches to detect interest in voting, and different approaches in weighing their data to reflect what they think the demographics of the electorate will be.
But pollsters don't have the final say. You do, by showing up to vote and by helping turn out the vote.
One of the reasons why Barack Obama outperformed the polls in 2012 was his campaign's ability to motivate unlikely young and people of colors voters. That should serve as a perennial reminder not to let polls dictate who participates in our elections.
This year, the polls may create a new X-factor by signaling that the outcome is a fait accompli. Hillary Clinton is well ahead of Donald Trump in national polls and well enough in most swing states. This has been the case all year, and is unlikely to change after two debates are behind us.
We haven't had an election in sometime that wasn't considered close since 1996. (While 2008 was never really close, many harbored fears that racist white voters weren't telling pollsters the truth. In fact, the 2008 polls were on the money.) And in the 1996 election, the polls created a complacency effect, altering the electorate.
The final Gallup poll in 1996 had Bill Clinton up by 11 points over Bob Dole. In the CNN poll, it was 16 points. Both had Clinton over the 50 percent mark.
But Democrats got complacent and didn't show up at full strength. Clinton still won, but only by 8.5 percentage points and with 49.2 percent of the vote. Coincidentally, or not, Democrats didn't make headway in Congress. They lost two Senate seats and gained a mere two House seats, leaving Republicans in control of both houses.
A similar dynamic could occur again, in which Democrats don't bother to vote because Hillary Clinton is considered a shoo-in. She would probably win anyway, but perhaps Democrats would not take back the Senate despite having several Republican seats in play. And a tighter race could impact public perceptions about her mandate.
But Donald Trump is not Bob Dole. There are already signs that Trump won't hold onto the entire Republican base, such a weak fundraising and several conservative editorial boards refusing to endorse him. (So far Trump has no endorsements from newspapers with circulations above 75,000 and one from the rest of the smaller papers.)
In turn, Trump's expected defeat might persuade Republicans skeptical about Trump to not bother and stay more. Moreover, Trump is repeatedly saying the election is rigged, which could make even his most loyal supporters stay home, thinking their votes won't be counted anyway.
If Republicans are the ones more likely to stay home than Democrats, then not only would Democrats have a better shot at taking the Senate, it could even put the House within reach. Or turnout could drop on both sides, leaving the Senate on a knife's edge.
Polls are a tool to help us understand the electorate, but we should not let them determine the electorate. No matter what the polls are telling you on Election Day, be the electorate. Show up.
