Brits settling down to watch the US election may well be expecting the evening to start with an exit poll – after all, that’s how UK election counts kick off, and they tend (with a few notable exceptions) to be pretty accurate.
Unfortunately, things aren’t so straightforward in the US, because each state has its own government and runs elections to its own set of rules, meaning the actual votes close at completely different times. (That, and the fact the US is split into six time zones.)
So exit polls in America are a more gradual process, but they can still tell us a lot about the result when we finally get them. And people will still comb over the early results for any hint of who’s likely to win. But just as in any other year, it’s worth using significant caution when looking at those early results.
Exit polls are a valuable source of data about who votes and why, but they’re also complex projects that have the potential to be misinterpreted, especially early on. This year, with an unprecedented level of early voting complicating the traditional process, that’s especially true.
For decades, news organisations’ exit polls have been conducted by Edison Research for the National Election Pool, a partnership now consisting of ABC News, CBS News, CNN and NBC News. Those interviews – a massive, nationwide undertaking – have been largely conducted outside polling places.
In-person interviews are still happening this year, albeit with a few pandemic-appropriate tweaks. Interviewers are masking up, leaving questionnaires on the table for voters to pick up themselves, and providing hand sanitiser, wipes and disposable pencils.
Taking those precautions, exit polls in this year’s primaries got similar response rates to those in past election cycles, Edison Research co-founder and executive vice president Joe Lenski said this summer.
Since 2004, in an attempt to account for early voters, the exit polls have also used telephone polls conducted over the final weekend of the campaign. Starting in the 2018 midterms, Edison began stationing interviewers at early voting centres as well.
Such methodological adjustments have gone into overdrive this year, as pollsters are faced with an unprecedented level of early voting amid a nationwide pandemic. This year, according to Edison, more than 25,000 early voters were polled by phone, and interviewers were deployed to early voting centres in eight states.
This year, there will also be an additional source of data – AP VoteCast, a newer, competing product being used by Fox News, The Wall Street Journal, NPR, Univision and PBS NewsHour. VoteCast, which was launched during the 2018 midterms, relies entirely on pre-election surveys conducted by mail, phone and online.
Exit polls, as we wrote back in 2018, serve a couple of different purposes. While they’re not foolproof, in the days to come, they’ll help provide a sense of how and when voters made up their minds.
News organisations also use them to help call the results of races (a task that’s also become exceptionally complicated this year, with profound disparities in how quickly states will count votes and a newfound partisan divide that’s seen Democrats far likelier to vote early).
Exit polls were not, however, designed to help give the American public a sneak peek of who’s likely to win, especially when people are still in the process of voting. Fundamentally, exit pollsare just surveys, and are subject to many of the same sources of error as any other poll.
And, like any poll, they need to be properly weighted to represent the population they’re supposed to be measuring. In this case, that means recalibrating the numbers to match the actual results of the election. Before that happens, exit polls can present a misleading picture of the race.
This year, with final tallies likely to take significant time, that’s more true than ever. As FiveThirtyEight’s Laura Bronner and Nathaniel Rakich explain: “The pandemic has undermined the one major advantage exit polls have over other kinds of polls: their ability to survey only actual voters. […] The phone poll essentially puts pollsters back in the uncomfortable territory of having to guess whether respondents who claim to have voted actually did.”
“It’s just really hard to know in advance what the proportion of in-person Election Day voters and absentee voters will be in 2020,” they added. “This election night especially, it’s best just to wait for the final results – even if it takes a while.”
🇺🇸 Running Mate is a US election podcast – just for Brits. Now available everywhere.