Politics

How the AP Will Decide When to Call States on Election Night—and When to Wait

A hand puts a ballot in a ballot box that has "I Voted Early" stickers on top of it
Early voting in Milwaukee on Tuesday. Kamil Krzaczynski/Getty Images

There is no formal mechanism for the United States to say, “We have a president.” Of course, there are 50 states that vote, and then there is the affirming of the Electoral College. But on election night or soon thereafter, we have generally defaulted to letting the television networks “make the call.” But in reality, the gold standard within the media is the AP, the Associated Press. Once the AP makes the call, well, then that’s it. The call is made. But as with all standards, this gold standard has changed, and it has had to change for this election.

As part of The Gist’s series “Calling It” about how the media is preparing for election night, I spoke with Julie Pace, who is the Washington bureau chief of the Associated Press. We discussed how the AP makes the call, and what it’s preparing for on Nov. 3 and the days—and possibly weeks—that follow. A portion of that interview is transcribed below. It has been edited and condensed for length and clarity.

Read Mike Pesca’s introduction to “Calling It” alongside excerpts from his interviews with NBC’s Steve Kornacki and CNN’s Brian Stelter.

Mike Pesca: What people are going to want to know is who won and when. You partnered with the University of Chicago to launch VoteCast in 2018, and you’ve been using it since then. That gives you a counting tool, but not enough to make the call. Could you take us through the portion of your task that is vote tabulation and research? How does that go down?

Julie Pace: VoteCast will be one tool, but it’s by far not the only tool that we’ll be using. We have a range of resources. We start with a group of analysts and race callers who are real experts in these states. Some people have been calling races in these states for several cycles. We draw on them pretty extensively here.

But essentially what they do is they are waiting to see if enough votes have come in, in the right places in those states to give us a margin [with which] we feel comfortable calling that race. That means that you could end up in a situation where we know that the margin [of victory] is 2 points right now, but we’ve got 5 percent of the votes still out, and it’s in those swingy areas of that state, a swingy county. It might mean, though, that if it’s a 2 percent margin and the Democrat is up, and all of the vote that is out still is from the most heavily Democratic county in the state, that might put us in position where we’re able to call that race. It’s just a series of data points that we’re looking at to make that determination. They don’t always cut one way or the other.

So it is the case that the AP will make a call even if it is mathematically possible for the candidate who you’re making the call against to win. The standard isn’t if Candidate A is leading by 20,000 votes, and if there are 21,000 votes out there, we simply will not make the call. No. You look at those 21,000 votes and say, “Oh, my God, they’re from areas that should heavily favor Candidate A. Therefore, we can make the call.”

Yeah. There are certainly situations where yes, based on raw vote that is still out [there] technically—if the trailing candidate captured every single vote that is still out, they could win. But if we know that that vote is from an area that votes 80–20 for the candidate that’s ahead, that will probably give us confidence to call that race. There are always caveats here. One thing that is always a really big caveat for us is we pay very close attention to what recanvass and recount rules are in these states. If we have a margin that is within that window, we’re not going to call a race, because we don’t want to have the AP race call be part of that process. We don’t want to be used in legal challenges or in the ways that states are evaluating whether to start that process or not. If the margin is within that window, we’re going to let that process go forward before we call a race. Again, always caveats. But yeah, absolutely. We’d look not just at what the raw tally is, but where those votes would be coming [from].

I read in the AP’s materials that you have a rule that you do not call a race if the AP determines that the margin of vote is 0.5 percent. You’re not going to make the call if you think it’s so close that you could be wrong.

That is typically our standard. Yeah. I think one of the reasons why we do that is because we only want to call these races once. We don’t want to call a race and then, say, learn that there is some change in calculation. We have had situations where a county can come back and say, “Actually, we didn’t put this number of votes into the system that we had,” not because anything nefarious happened there, but [because] these are human processes. We have found that sometimes there are just things that change. When the margin is that close, we don’t want to have gone out and called a race and then have to take that race call back.

That’s, I think, really central in this election. We want people to feel really confident that when we call a race, that it is definitive, that you can trust us, that this is not going to change. You’re not going to learn something else that happened, and we’re going to have to reverse that call. But that means that we have to have some level of caution within our race call to allow us to be certain that when we’re doing it, that it’s right.

Would it surprise you if there is a situation where the networks have a bunch of check marks next to states they’ve called, but the AP doesn’t? Maybe four, five, six swing states where other big media organizations with lots of audience have weighed in, but the AP is holding back?

I would say, if you look historically, you’ll see a lot of examples where we’ve called races that other outlets haven’t called, or they’ve called races that we haven’t called yet. We all take a similar approach, but we all have tweaks around the edges on this stuff. It’s certainly possible that, yeah, they could call some races before we’re ready. It’s also possible that we’ll call races before they’re ready. We were in a situation in ’16 where there were some states, individual states, where we hadn’t called the race yet when some other outlets had, but we ultimately called Wisconsin and then the presidency before others did.

Right.

Sometimes it kind of balances out at the end. We’re just going to kind of play our own game. That’s how we’ve always done it. We were very clear in our newsroom on election days: We care about the AP race count. That’s the one that we control. It’s the one that we believe is the gold standard. It’s the one we want to point people to and make people feel confident in.

So far, we’ve been talking about accuracy and numbers, a very important part of journalism, but an essential part is narrative. What will you, as an organization, maybe even you personally, be doing in your explanatory or storytelling capacity to explain to the voters, the audience, Americans, and maybe temper expectations about what election night really means?

We have been writing for some time, since the summer really, stories trying to lay out for people this simple idea that you might not know the winner on election night, and that does not mean that something has gone wrong. That is just a function of the fact that we’re expecting more mail-in vote this year in some states, including a state like Pennsylvania, which is a really important state, isn’t going to be able to start counting those mail-in ballots until pretty darn late. That could really slow down that process. We’ve been trying to really just hammer home this idea that it’s OK if there’s no winner on election night. That’s not going to probably mean that there was some kind of widespread fraud here.

The other thing, though, that I think you’re going to see from us is I think we have to go above and beyond this cycle in explaining the race calls that we are making and explaining the race calls that we’re not making. I don’t think it’s going to be enough anymore in this environment—as much as I think that we’re the gold standard—I don’t think it’s enough in this environment to just say, “AP has called this race for this candidate,” or “AP is not able to call this race at this time,” and expect that people will just take us at our word.

Stay tuned for Wednesday’s episode of The Gist’s special series “Calling It,” which will feature an interview about election night misinformation with Brian Rosenwald, a historian and author of Talk Radio’s America.

Listen to this full episode below. To get The Gist every weekday, including this election series, subscribe on Apple Podcasts or wherever you get your podcastsJoin Slate Plus and enjoy ad-free episodes of the show.