Decision desk editors at The Associated Press, Fox News, and CNN joined PEN America on Wednesday, October 14 for a lunchtime webinar to discuss election calls, projections, and how news organizations are planning to navigate the challenges of election night 2020. Moderated by broadcast journalist Ray Suarez, this conversation included the following panelists:

  • Sally Buzbee, Senior Vice President and Executive Editor, The Associated Press
  • David Scott, Deputy Managing Editor, The Associated Press
  • Arnon Mishkin, Decision Desk Director, Fox News
  • Dana Blanton, Head of Public Opinion Research, Fox News
  • Sam Feist, Washington Bureau Chief and Senior Vice President, CNN
  • Jennifer Agiesta, Politics Polling Director, CNN

Check out the webinar recording below or read our transcript to learn more about how news organizations will or won’t make calls, what responsibility networks and news outlets have in preventing the spread of misinformation, and how election night may stretch into election week.

The conversation is part of PEN America’s What To Expect When You’re Electing initiative, designed to inoculate Americans against the threat disinformation poses to this year’s election and to democracy.

SUZANNE NOSSEL: Hi, my name is Suzanne Nossel. I’m the CEO of PEN America. The mission of PEN America is to both celebrate and defend freedom of expression here in the U.S. and around the world. Back in 2017, we began to examine disinformation and concluded that even though propaganda and fraudulent news are by and large protected by the First Amendment, they nonetheless pose a serious threat to open discourse.

Like insider trading and information asymmetries in financial markets, disinformation represents a market failure in the bazaar of ideas, standing in the way of our ability to persuade one another based on the merits of our arguments and awarding the power of truth to rise to the surface. When it comes to the upcoming election, disinformation, conspiracy theories, and the deliberate sewing of chaos and divisions are rattling the core of our democracy.

That’s why PEN America launched What to Expect When You’re Electing, a national initiative aimed at inoculating voters against disinformation, by ensuring that people know what they can anticipate in this election season like no other and have the tools they need to avoid trafficking and disinformation. The campaign involves town halls in swing states, a fantastic video featuring John Lithgow, Alan Cumming, Anita Hill and others, tip sheets, the quiz and training program, with partners all over the country.

As part of that effort, I’m delighted to introduce today’s discussion on decision desk with CNN, the AP, and Fox News. Just as the process of voting in this election will be different, so will the work of news organizations that we rely on to convey results. We believe that the more Americans know about how the networks gather information, make their projections, and deal with the forces trying to shape how the selection comes out, the better informed and discerning we can be on what might turn out to be a very long election night or even an election week or month.

I’m delighted to introduce Ray Suarez, who served as senior correspondent for the PBS NewsHour for 14 years and has reported for NPR, CNN, ABC, and CBS. When we thought about who could lead a conversation about how to guarantee credible coverage at this pivotal, perilous moment, we thought immediately of him. Over to you, Ray.

RAY SUAREZ: Well, thank you. And it’s really fun, after a lifetime of getting marching orders, Election Day editorial meetings, saluting smartly, and carrying out the policies I’ve been given to moderate this panel of insiders, who are ready to take us for a glimpse behind the curtain to reveal the thinking behind the decisions their institutions have made about how they’ll handle this unusual nationwide ballot, what they anticipate, the pains they’ve taken, and the plans they’ve made about the foreseen and unforeseen challenges of heading to the polls in the midst of a pandemic and in the midst of a presidency that’s challenged the norms of reporting in so many ways. I’m joined by Sally Buzbee, senior vice president and executive editor for The Associated Press; David Scott, deputy managing editor of The Associated Press; Dana Blanton, head of public opinion, research for Fox News; Arnon Mishkin, the decision desk director for Fox News; Sam Feist, the Washington bureau chief and senior vice president for CNN; and Jenn Agiesta, the politics polling director for CNN.

There’s well over a century of accumulated experience in this conversation, and innumerable election days. If you’ve been thinking about how the business will handle November 3rd, you’re exactly in the right place. One of our panelists noted that our system here in the U.S. for finding things out about elections is unusual. We don’t just wait for the results from the government, but from the news media, and that the pressure to get it right, to be credible, maintaining the public’s faith in our system is part of the explicit and implicit assignment. So getting it right is important, not just because that’s what we’re always trying to do, but there’s a profound civic responsibility here and threat of misinformation. It’s a big panel, and since we’re opening with brief initial presentations about election night policy, what you’ve all decided about how to handle returns and reports, I would only ask as we begin that you keep it tight, brief, elegant, and complete.

Sally, for decades, AP has been a go-to source for numbers and results—what have you got planned? What did you have to consider in that planning?

SALLY BUZBEE: So, one of the things that I think is important to keep in mind for this year is that it’s not really different in terms of how America fundamentally conducts elections, or even fundamentally how we call races. It is in line with the past—America conducts very decentralized elections. Each state runs its own election, and then those numbers are tabulated. But this year, we do have to apply things that were really only happening in a handful of states, now too many states. That, to me, is one of the biggest differences.

So, for example, in the last presidential election, about 40 percent of Americans voted before Election Day—either absentee or mail-in ballots or early voting. Accelerated by the pandemic this year, there’s about two decades probably worth the change that are happening in one year. And so, this year, most of the expectations are that 55 to 60 percent of Americans could be advanced voters of some sort this year. Five states had mail-in ballots for many years and handled it quite effectively.

But this year, there are just many more states that are here handling mail-in ballots. What the pandemic has sort of done, through the addition of advanced voting, is increase the scale of where we might need to handle these things. There’s nothing that’s really new, but the scale is different. So that has complicated our jobs in some ways, because there are now more places doing that. None of this was new—we have experience and states have experience dealing with mail-in ballots, but we’d have to apply that experience to more places. So, for example, if a state is somewhat slower to count its mail-in ballots, which we find often does happen, then we need to adjust our race calling to make sure that we have an accurate view of all the outstanding votes before we call a race.

That’s one of the major changes. My colleague David Scott is going to talk a little bit about one of the other changes that we see this year, which is really just a renewed commitment by us. And I know many of them are news organizations providing greater transparency about our methodology and how we call races.

DAVID SCOTT: Thank you both. When we call a race, the AP in the past, it’s 120 characters. “Joe Biden wins Delaware.” “Donald Trump wins Florida.” In a lot of ways, in the past, that was sort of the headline, and then we would go on and tell the rest of the story. We found that from conversations with customers, and that’s our role as the AP, as a news agency, but also with our customer’s customers, readers, and viewers. They want more than that headline now. They want to know how the call was made. What went into our thinking process? Why was that moment the moment that could make a race call? And what went into it?

Some of our customers who rely on AP for our calls, they want all the same information that we use to make race calls so they can essentially check our work. Being transparent and showing our work and showing how we make decisions and the analysis and the logic that led to a race call is something that we really started to formally incorporate into our race coverage in 2018, and it’s so much of our coverage and will be this year.

We need to sort of fight misinformation and do a lot of explanatory journalism around the act of just covering the story. That’s another thing that’s not so much different, but just on a larger scale this year is more transparency, more showing of our work about what goes into a race call.

SUAREZ: Great. I’ll move on to Sam Feist from CNN. Sam.

SAM FEIST: First of all, thanks PEN America for pulling this all together—it’s a really timely event. I want to start by saying that I’d like to affiliate myself to everything that Sally and David just said. The AP has been calling elections for a long time, and I think you will find across news organizations that call elections, the AP and all the television networks that are decision teams, they all know each other.

They have been doing this for a long time, and they use remarkably similar methods and techniques to check and double check the races before we make projections. As you were asking, Ray, what are we doing to prepare? I think the most important thing that we can do as news organizations and as leaders of news organizations is to prepare both our audience—in this case, the American people—and our news organizations for a somewhat different kind of election night. And the word that I am using over and over is patience. This could be a very orderly election, whereas Sally said the projections or the vote count can happen in a very orderly fashion. They may take a little bit longer because it simply takes longer to count mail-in ballots—you have to open an envelope and then another envelope and check a signature and unfold the paper and process it. And that takes a lot longer than an in-person ballot today does to count.

But if it takes longer, that doesn’t mean that there’s anything wrong. It just means that it might take longer. So we all have, I think, an obligation to remind people of the processes, to be transparent about how we do what we do and educate our audiences about how votes are counted, why it might take a little bit longer in this cycle than it has in other cycles. If there’s not a winner on election night, that doesn’t mean anything is wrong. There may not be a winner on Wednesday morning. There may be. That doesn’t suggest that there’s any problem whatsoever in the system. But we need to help folks understand that. And the other thing I would say as we start this conversation with several different news organizations, I think it’s important to note that there really is not a competition between us to be first on a state or first to declare a president, and that’s really important.

It may have been different in 2000 between news organizations and in the race. But I will tell you that an awful lot has changed since 2000, in the way we handle decisions and reporting of results. I can assure you—I can speak on behalf of CNN, but I suspect I’m speaking on behalf of others, that we are not competing with each other. The numbers are going to dictate when we make a projection. Sometimes, projections come at a similar time by multiple news organizations, and that’s because the numbers have changed and have set us up to make a projection. Other times, AP may be ahead in this state, or CNN may be ahead in this state, or another network may be ahead in this state. We’re not racing each other. We are not competing. And I think that’s a really important piece of this conversation. 

SUAREZ: Jenn, let me go to you next. Then, then we’ll head back to the Fox Team.

JENNIFER AGIESTA: I think the major thing that I would like to add to what Sam said, and I’m so glad that you talked about 2000, because Sam asks me and my entire decision team to remind ourselves exactly what happened in 2000 before every major election night, so that is a big focus within our organization.

What I would like to add to that is we’re going to have this willingly large increase in absentee and early voting happening this year. But we are also in a great position to know a lot about what that vote is and what it means, heading into election night. The unknown, really, is how Election Day itself is going to go.

We don’t know how many people beyond those who we know have returned ballots in reelection by mail, or who’ve turned out to vote early. We don’t know how many people beyond that are going to vote on Election Day. We don’t know where there might be a coronavirus outbreak that could affect turnout.

I think until we get a good sense from the numbers and from the data on what that Election Day vote and turnout looks like, I think you’re going to see all the decision desks, including CNN, being very cautious in what we do through election night. But I do think it’s important to remind people that we know a lot of things about what we’re seeing in the absentee and early vote and about Election Day voters from our exit polls. So, we have a lot of pieces of information that we’re pulling together, and we will get there. It just may take a little bit longer than it typically does.

SUAREZ: Here’s Dana Blanton.

DANA BLANTON: I want to piggyback off of something that Sam was saying about not seeking a competitive advantage. I think that when we left the consortium in 2017 to come up with something new, it was truly to just improve accuracy, and it wasn’t to seek a competitive advantage. As we developed a new idea to cover the election, we actually met with all the members of the NEP—and other potential partners—when we were trying to come up with something new. This something new is that we use the same methodology for all voters—early voters, Election Day voters. That’s why we feel just so well-positioned to cover this election, and especially to talk about the difference between the early and Election Day voters.

This is something that we saw clearly in one of the very first elections that we covered with this new methodology—when we were testing it in 2017 with the Alabama Senate election. We saw how Doug Jones lost to Roy Moore on Election Day, among Election Day voters. He was able to win the race because he had in the bank so much from the people who had cast their ballot early. So, in addition to all of the other things that we typically get from the massive election survey, we’ll be able to talk about that difference as well as about people who choose not to vote. We talked to non-voters, and that can sometimes make a difference in an election. This year, in addition to the regular reasons that people might not vote, people might be concerned about coronavirus, or maybe they’re confused about how to cast their ballot. So, we go beyond just looking at the early and Election Day voters, and also talk to the non-voters. I’ll just say that as much as we enjoyed working with the consortium—and I really did love working with the consortium—the public benefits from having two voter X-rays.

ARNON MUSHKIN: I assume you want me to go on now. Just to continue to build on what Dana said—and also to build on what Sam and Jenn said, as well as Sally and David—there isn’t a competition amongst the networks in terms of being first. At the same time, we’re all competitive animals, and there is a competition in terms of being able to tell the story. I think that we’re all going to work very hard, and I’m speaking just for Fox News. In terms of trying to be as accurate as possible in telling the story of this election, which—regardless of whether or not you’re looking at polling from Fox News or from CNN or from the AP, or various other places—everyone is in agreement that this is viewed as the most important election in many people’s lifetimes. So, I think this will be a high-visibility election on which there will be a competition to try to tell the story as accurately as possible.

Clearly, one of the biggest stories that night is going to be the difference between mail-in voter and Election Day voter—mail-in vote and in-person vote. One of the things that that has told us is that our traditional models for analyzing an election, the historical models don’t work. The reason for that is, it doesn’t matter whether you call your model a county model or an integrated model or a needle or anything else—they’re all built on one core assumption, which is that when you start getting the vote in a particular county, from the get-go you have a good picture of what that vote is going to look like. So, once you get 20 percent of Raleigh County—or, frankly, once you get two precincts in Raleigh County, West Virginia—you multiply that by the number of precincts in Raleigh County, and you get a good estimate of what Raleigh County is. That’s true for about all the 4,000 jurisdictions, with a few exceptions that those of us who work on election nights, election desks, know.

What we realized this year is that the assumption that’s built into all the models that we’ve built over the past several decades doesn’t work. When you start seeing a vote, you don’t know if you’re looking at a typical precinct, you don’t know if you’re looking at a mail-in vote, you don’t know if you’re looking at an Election Day vote. From all the polling, we know that if you’re looking at mail-in vote, whatever numbers you’re seeing are probably skewing to the Democrats, from where the average is. If you’re looking at Election Day votes, those numbers are probably skewing to the Republicans, given what everyone has told pollsters up to now. So, we had to rebuild our models around that realization. Dana talked about how we’re working with the AP on what we call the Fox News Voter Analysis and what they call VoteCast. We think these tools will help us to get a really accurate idea of what America is saying, what voters are saying, and how they’re saying it, and what it means. I think that they’ll be well-served not just by Fox, but also by all the other networks on election night.

SUAREZ: Jenn, earlier you talked about the need to keep up with events in order to understand what you were dealing with on Election Day itself. As early voting opened in some key states, there are huge early votes in-person, and also large bodies of votes already taken in states like Florida and North Carolina by mail. How does that—as it’s building over the next two weeks—change how you’re anticipating what needs to be done on Election Day?

AGIESTA: I think it helps us to set expectations in some ways around turnout. We know, for instance, in our CNN count at the moment, there’ve been almost 13 million votes cast already in the United States, across 42 different states that are counting, or that are reporting those numbers right now. That’s roughly nine percent of the total votes cast in 2016. We’re going to be watching that percentage over time and seeing how it changes. We can also look at—as those ballots are returned and secretaries of state start to count voter records—we know a good amount about the partisan makeup of the people who have returned their ballots versus those who requested them and have not yet returned them. We’ll also know a little bit about their demographic information—about their gender, their race, their age, whether they voted in the past. And those kinds of things are clues about what turnout might look like in this election.

Now, I do think—given what we’ve been seeing in polls around enthusiasm for voting, interest in this election, combined with the massive turnout spike we saw in 2018—it’s reasonable for us to expect that our turnout is going to be massively higher this year. The tricky thing for us to figure out on election night is going to be whether the increase that we’re inevitably going to see in absentee and early, how much of that is the increase in new voters, in turnout, relative to 2016 or 2012, and how much of that is sort of cannibalizing the Election Day vote and reducing that number. I think that being able to answer that question, and having enough data to answer that question, is going to be just the most critical thing we can figure out on election night.

SUAREZ: Sally, with all the variables that our various panelists have mentioned in the last 10 minutes, have you been able to throw extra troops? Are there bodies to cover contingencies, to bone up on election law, to understand the difference of the count in the mail-in ballots in one state as opposed to another? We’re operating in an environment where the business model is under a lot of pressure, and a lot of places are not as well-resourced as they used to be. Do you feel sure now, inside three weeks, that you’ve got what you need?

BUZBEE: We started to answer that. One is that we, like the other folks on this call, have a group of core people who are highly skilled at race calling, highly skilled at race analysis. We haven’t necessarily added to that crew of people. We’ve trained hard, as we do before each election, but we have our core, most analytical people who are excellent at race calling. What we have added, though, is—and I think the way to explain this is that the state of election law in a state like Pennsylvania or Texas is still in some confusion right now, there are still ongoing lawsuits, and things like that. So, it’s critically important, obviously, that we understand what each state’s election law actually looks like by the time people start counting votes, by the time the states are counting votes. How long will they take absentee ballots? They have to be postmarked by Election Day, but if they get there five days afterward, are they going to count them? So, we have a lot of research effort.

I guess I would say, reporting by our power that we have thrown at this year, to make sure we’re on top of the election lawsuits, to make sure we understand what the laws are doing. That has been a lot of work. We calculate the vote on election nights, so we have somewhere of like 6,000 stringers on election night that are helping us gather the vote. What we’ve done this year is just add a lot of reporters. We also need to be on top of it, just from a reporting point of view, not just for lawsuits, but also if there are protests or disruptions or voter intimidation that are going on. Those are all part of what we’re going to need to be able to report on before Election Day, on Election Day, after Election Day. So we have added people, resources to that effort most definitely.

I also just want to say one thing. What we try to do is we try to be as accurate as we can. I just want to just really reinforce what Sam and Jenn said. We are competing to be as accurate as humanly possible, but we are really trying to address our mission and not compete with other people in terms of being first. When I hear Sam say that, I am deeply reassured about the state of the media in the United States, that everyone is really trying to do that most ethical thing. And I admire everyone on this call so much because I know just how ethical and fact-based they are in their approach to this. And I just think that that transparency is enormously important to make sure that the audience also knows that it’s happening.

SUAREZ: Dana, do you come into this cycle much more conscious of the educative role? Not that you want to make your election night coverage into a freshmen poli-sci seminar or something, but there’s a lot that has to be explained. There’s a lot that has to be untangled along with just the normal tabulating and explanation of what we know so far and what’s been able to be counted so far.

BLANTON: I think that’s right. I think that it’s both educating the talent and the anchors who will be covering election night, as well as educating the viewers. I mean, we’ve been talking and talking about, it might not be that night. It might take us several days and to let people know, and including even asking questions on our regular national Fox survey about, when do you think we’ll know? So that gives us another opportunity to talk about it so that you can say, “Well, some people think we’ll know that night, but a lot of people know then it’ll take possibly a few days.” We’ve also worked on creative and new ways this year to talk about how we can show people what’s happening, new graphics to try and show. Arnon’s working out the models to show this is what we do know, and this is what’s outstanding, and what does that mean? What does that tell us about what might happen in this state?

MISHKIN: One thing just to build on, in terms of the educator role—there’s a lot of talk about how this country is very divided. We live in a very divided country. In reality though, very few of us live in divided communities. In fact, most of us live in communities where everyone is going to vote, live the way we do. And that’s true, whether you live in a blue state or a red state. And I think that one of the important roles on election night is to try to make the case or describe for people the fact that we do, in fact, live in a very divided country and to show people where those divisions are, because we often don’t see that when we walk into a grocery store.

SUAREZ: Excellent point. I don’t know whether you all saw the piece by Mike Murphy in The Washington Post—a famed political consultant, now a podcaster. He used the lovely phrase, “malicious doubt merchants,” which I’ll get to later. But he suggested that the exit polling should not just reflect people who are coming out of polling places on Election Day, but the design of the sample should be changed in a way that reflects early voters and mail-in voters. If you haven’t done that already, is it too late, with just a few weeks to go? It sounds like a smart idea, but is it easier said than done?, When we’ve been talking so far about how much of the vote is going to come in before the polls actually open on November 3rd.

FEIST: Let me try to tackle that. Even before the pandemic, last year, years ago, we have always taken into account the fact that many people vote before Election Day. And so, I appreciate Mike Murphy’s suggestion, but this is well baked in the national election pool. I think Arnon mentioned the consortium is a consortium of ABC, NBC, CBS, and CNN.

We have two mechanisms for surveying the voters. We have in-person exit polls, something that we’ve been doing for many years, both on Election Day, but also we’re doing it right now. We actually have exit pollsters who are out at polling stations across the country that are open and are surveying voters to find out how they voted.

But far more important this year than in any other year is the phone poll survey of voters who may be voting by mail, who may have voted early—tens of thousands of those surveys are happening. Obviously The Associated Press will do this, we’ll do the same thing. We are surveying early voters. It is a critically important part of the electorate. Maybe Sally’s numbers turn out to be correct. It may be the majority of the electorate. We’ll find out eventually, but yes, that is an enormously important piece of it. So yes, we will know in advance, election night, the views of early voters and the views of today’s voters. And it’s also important that this is a part of our conversation on election night. It’s a part of the transparency that everybody’s talking about.

We’re all adding into the conversation—something that has been less important in other years. And that is the conversation about how the early voters and the mail-in voters voted versus how the today’s voters voted because, as has been noted, they’re not necessarily the same. We see in pre-election polls that Democrats have a much greater propensity to vote by mail than Republicans. That’s important for us. And that was a part of Mike Murphy’s column this week. We have to help our viewers understand that. So, on CNN, as we report vote results, you’ll see John King at the magic wall that has been an icon of our election coverage for many years. We will show where the votes are coming in geographically. But we will also, this year—more than we have ever done—talk about what we know about the today voters and what we know about the mail-in voters. And of course, many of those mail-in voters won’t be counted tonight, and we have to explain that and help the viewers understand that there’s more to come, and that may be later tonight or tomorrow or even the next day, because it just takes longer to open those.

SUAREZ: So many of you have mentioned the word transparency in this conversation so far. And I want to talk a little bit more about why that’s so important, not just covering a difficult, challenging, novel Election Day, but also the need to be frank with your listeners and readers and viewers about what you’re saying, why you’re saying it, and how you got there. Sally mentioned the phrase, “Show your work,” which was the nightmare of math homework for years and years in school. Why did so many of you mention transparency?

SCOTT: I think, for one reason, I mean—Jenn referenced 2000 before. I’m thinking about how many sources of information there were in 2000—for news about the election—and think about where we all got our news about politics 20 years ago, and where we get it today. And the barriers to entry for people who want to be part of that conversation, as a broadcaster, as a headline maker, as an originator of content—I mean, obviously there’s been an explosion, right? And it’s so much easier to enter that conversation now than it was in the past. And so, one of the ways that groups like AP and CNN and Fox can differentiate themselves in that marketplace, is to show that work. You don’t question the partisanship, for example, if you have the math behind it. What are the intentions of someone who might be on social media or what are their goals? It’s harder to make those questions if the math is there, if you’re showing your work. This is how we reach this conclusion.

This may not be the answer that you expected to come out of this state. Certainly, I think I can say for most everyone on this panel, Wisconsin wasn’t on the top list four years ago. But so when something does happen, that’s unexpected. You can show the math, you can show the work, you can show the votes, you can show the research, you can show the survey data that illustrates how you arrived at that answer, how you arrived at that call. And not everyone in that marketplace for news can do that. And that’s the value of it. We’ve made the investment in the research. CNN and Fox and the other networks make the investment in the decision desk. We go out and tabulate the vote, we conduct these surveys, and that has value. And again, showing that work—I think it adds to the power of what we do on election night. Just that transparency is just so crucial, now and moving forward.

SUAREZ: I’m glad you mentioned the number of sources because Charlie Kirk, just to pick a name, has more followers on Twitter that the circulation of any newspaper in the country and in that sort of situation where he is an opinion maker and an impression maker, you no longer have the field open to yourselves as the stamp of credibility, the stamp of approval, the take-it-to-the-bank sort of status that you once had, through no sin of your own. But because there are so many people who have crowded into the arena and they’re able to shout their opinions out into the air and inject their opinions, sort of into the public bloodstream, it is a transformed landscape. That means the news business has to, in effect, prove why it’s the news business while everyone’s watching.

SCOTT: I think that speaks to my point though. I mean, first off, I’d say half the world’s population sees news from AP every day. And I’m sure Sam and Arnon and Dana and Jenn have some stats like that too, that our reach extends beyond just social media, but that’s the thing that we can do that someone like the person that you cited can’t do. You know, what’s behind your tweet? What’s behind your race call? I know that Jenn can point to the data that illustrates why she called the state. And I can expect that in the very few minutes that I’m going to have not focused on my screen in front of me on election night, if I was to flip over to CNN, she’ll be explaining it. Same thing that Arnon and Dana are going to do that data behind it, that explanatory behind it. That’s the difference maker. And I think that’s the most powerful thing that we can do on election night.

SUAREZ: Will you all, in your ways, go into—whether you intend to or not—creating an impression about how this Election Day went? If you go into this process already signaling to the audience to expect trouble, to expect confusion, to expect delays, does that create one expectation in the audience, long before we’re even in a position to know anything, versus saying, “We’ll tell you what we know when we know it, have patience”? Sam, does your tone, long before there’s a single result to announce, long before a single poll closes, create in the public certain expectations?

FEIST: Sure. Our tone is important, and explaining what we know and how we know it is important. And as I had mentioned earlier, a delay is not necessarily a problem there. You know, people forget that it wasn’t just 2000. In the 2004 election, we all went to sleep without having projected a winner. In the 2004 election, most news organizations had not yet projected Ohio, and Ohio was the state—whether Kerry or Bush won Ohio would determine the outcome of that race. That didn’t happen until almost noon the next day. People forget about that. So, not necessarily knowing a winner on election night is not reflective of a problem.

And I just want to underscore that, obviously there are always issues on election night, every election night. There may be a polling station that loses power. There may be a weather issue. People can’t get to the polls. Some polls, there may be long, long lines and, an election administrator or a judge may keep a polling station open longer. Those are not necessarily problems, and they may have no material effect on the outcome of the election in any way. They are frequently worth reporting and frequently not worth reporting because they are insignificant. There are always going to be issues. We have scores of reporters, fanning out across the country now to report on various voting issues or legal challenges such as the back and forth challenges over the Texas drop boxes. The governor said one box per county, a judge overruled that, and an appeals court overruled that. Well, that’s something that we have to cover that doesn’t reflect that there’s a problem in the system, but it’s worth us explaining to the viewers. It will also explain how people in Texas are going to vote. But remember, if the election is not decided on election night, if it’s not decided perhaps until Wednesday or even Thursday, that is not necessarily a problem. That’s just counting votes, reporting them as they come in, taking as much time as we need to make a decision. And none of us are in a rush.

SUAREZ: There are more than 250 people in the room listening to our conversation and we’ll get to questions in just a moment. But before I do, I want to ask Arnon and Dana about the internal conversations at Fox leading up to Election Day. Because you have opinion shows and news shows, is there a necessity for sort of “rules of the road” as you head into a day as critical as Election Day, where you work out sort of the body English between the news side of the house and the opinion side of the house?

BLANTON: Well, our election night coverage is a news division that’s run by the news division. And so, I think that that’s how it’s covered, and that’s how we’ve always done it in the past. I think I want to also go back and talk and just add to something that you were talking about earlier that David was talking about as far as the transparency, in addition to the kind of things that we’ll do on air. We post all of the information that we gather: all of the surveys, the 50 states, the national—we post all of that information on our website at poll close time. And so, we try and give people all of the information that we can. And that’s just another way that we’re being transparent and giving people the information: what we know when, at the same time, right? And so, people can go and find that for themselves and keep up with it, continuously updating, and we provide all of that.

SUAREZ: Remington asks, “How do you prepare against the disinformation of fake news through bots? In the 2016 election, bots had a significant impact in the act of spreading fake news and spreading disinformation. How do you approach that as a media organization? How do you approach Americans who have been convinced that these fake news platforms are the real news?” Who’d like to handle that? Go ahead, Arnon.

MISHKIN: Our job on election night is to report the counting of the vote, which is not to report on what even either actual opinion people, or for that matter, bots, are saying about the vote, but rather what the voters are saying in their votes. And that’s the sole focus of the election. I think NBC is not represented here, but the old NBC president—I think Rubin Frank—said, “It’s a TV show about addition,” meaning it was going to be quite boring, but on one level, this is a cable TV and now internet show about addition. And I think the way we approach that, at least from a decision desk and polling desk perspective, is let’s just focus on what the voters are saying, and what they’re doing in terms of their ballots.

FEIST: And Ray, this is just a good opportunity to make a plug for mainstream media. If there’s ever a time to rely on trusted sources of news, it’s around an election. There’s going to be a lot of noise out on social media, but the vote counts are coming from two places: the national election pool consortium and The Associated Press. This is where the vote counts coming from any other source of the vote count nationally may be suspect. And really, elections more than any other time is a moment when people need to turn to news sources that are held accountable for providing reliable sources of information. And that’s what we will all do on election night. And we’re going to work very hard, we’re going to spend a lot of money, dedicate a tremendous number of resources. None of the social media noise makers out there have the resources to actually do what we do, which is survey the public and tabulate the votes from tens of thousands of voting precincts, communities, townships, counties across the country. And so, go to your trusted sources of news, and you’ll get your information on election night, and it’ll be accurate.

SUAREZ: Elliot wants to know, “Is it your job to ignore what the campaigns are saying on election night? In other words, are you only looking at the numbers and not taking into account what people on the ground are saying about turnout, voting problems, intimidation, etc?”

BUZBEE: I mean, our race calling is not dependent on—let’s say someone from a campaign calls us up and says, “Hey, we know we won!” you know? We discount that. And in fact, I think I can speak for us—but I think this is probably true of everyone—we keep our race callers in a little bit of a bubble so that they are only looking at the factual information and they’re not paying attention to some of the other things that are going on. The only caveat I would say to that is that, from a reporting perspective, the campaigns are trying to find out facts really fast too.

They may be outwardly spinning, but they may actually know some very factual information. So, we obviously talked to them, but we don’t trust anything they say on election night. We put everything through our normal reporting process. If someone calls me up and says, “There’s a big disturbance at this polling place,” what we will do is send our own reporters there for eyewitness accounts to see if in fact this tip that we ever received is true or not. So, we have enormous—it’s almost inconceivable—every single person on this call. We have enormous reporting cores, whether they’re doing vote count, whether they’re doing what you think of is reporting. We are out there, and we are looking at what is happening, and we rely very strongly on facts and firsthand information. They always try to spin it. That’s like an old story. We all can handle that, we don’t listen to them.

SUAREZ: David, briefly.

SCOTT: I was just going to say that I like all sources of data. We’ll look at everything and we judge it, but at the same time, to go back to something that Arnon said about the election night show being about addition. It’s hard to spend sums, right? You can give me all the lines in the world about how your guy’s going to win, but two plus two is always going to equal four. There’s not a whole lot of value on the decision desk anywhere trying to listen to that, because the math is going to catch up with you. It’s all about the math, and that’s not going to change no matter how great your line is, or how much pressure you put on us, or how many texts you try to send us. The math is the math.

SUAREZ: Next is David, who asks, “Would you all consider agreeing to project simultaneously using the same agreed upon standards? Why compete to call states first when it’s so potentially dangerous?” Well, you may note David, we touched on that earlier, but let’s talk about how the competitive pressures actually manifest on a night like election night and why that hasn’t happened—what David’s suggesting.

FEIST: I think we compete on election night in different ways. We compete on—at least in the television networks, we compete on presentation, we compete on approach, we compete on our anchors, our analysts, the folks that we have on our air, the methods by which we report information. Some networks have more conversation. Some might be more focused on the data or the numbers. Some might be in the middle, but we are—I said it earlier and I really believe it—we are not in a race. That would be counterproductive for all of us. We know that we have to wait for the numbers to come in. Nobody’s in a race to jump the gun, nobody’s in a race to beat another news organization on a state. It just doesn’t work that way. And I think that’s a good thing. It’s also a good thing that we have two very distinct sources of information. There are two very large organizations counting the vote. Two very large organizations that are serving the public in the form of voter surveys or exit polls, and they, in some ways, provide a balance and a check on each other, and that’s a good thing. That’s something that we actually embrace, and I think is beneficial for the electric, for the public, and for the country.

SUAREZ: Kevin wants to know in a close race, “Is there uncertainty in making the call this year arising from the track record regarding the percentage of absentee ballots rejected from states new to absentee balloting?” Jenn, do you want to take that one?

AGIESTA: Sure. I do think that that’s going to be a concern in close races, and I think that you will see sort of what our normal threshold for a close race might be a little bit larger than it has been in the past. There are a lot of possibilities for legal challenges—if we’re talking about a very small margin between two candidates and there is a long time after Election Day during which several critical states will still be accepting incoming ballots. So, if we’re looking at a race that is a point or two points, I don’t know that we’re going to be able to make any projections in those places, because we’re just not going to have enough information about what the full range of things to be counted are, and the rejection rate is really just a piece of that. And it’s certainly something that we’ll consider particularly in states where there’s been a huge increase in mail balloting relative to what’s happened in the past. Some states are very good at this and have been doing it this way for a really long time, and so I’m less concerned about that in those places. But the places that have expanded pretty significantly this year, we’re going to let the election officials do their jobs in those places and count those votes and figure it out before we make any projection in a race that’s close.

SUAREZ: When you say “still accepting ballots,” they have no way of knowing how many ballots are still outstanding and how many are eventually going to come in, right?

AGIESTA: Generally, most states do keep account of ballots requested, so we’ll have some sense of how, in terms of many ballots that are out there. Not all of them will come back—obviously that’s your max ceiling of how much could be out there, and they need to be postmarked in time, and all of those other things need to be met. So, there’s a lot of uncertainty about it, but we can say this is the most that we would expect from most states.

BUZBEE: But I think this is a really important point. And I’m not as knowledgeable about this as Jenn or David, but from a layperson’s perspective, if there is any ambiguity, if state election officials have some ambiguity about how many outstanding votes there may be, and it’s a close race, obviously that throws up a caution note. We have to know, if it’s a close race, how many votes have not been counted yet and released to be able to make a race call. So the question is the amount of a very specific thing. This is actually one of the major sources of the race calling potentially being slower this year.

SUAREZ: This one’s from Tyler: “How will news organizations differentiate and stress the difference between too early to call and too close to call?” And I guess that piggybacks on what we were just talking about—we don’t know what too close to call is if we don’t know what the entire universe of ballots is going to be in, in mathematical terms, right? Fox team, why don’t you take this one?

MISHKIN: So we do differentiate, and early in the night—if once polls are closed, based on what the Fox News voter analysis is showing—we characterize a race as tilting one way or the other, or we may characterize the race as being too early to call it. Most often this year, it’ll be too early to call. It doesn’t really become too close to call until it really is, “We had a lot of data, and we’re waiting for even more data, and it’s right now too close to call,” or the uncertainty is too great. And we’re going to try to dig deep into explaining what that uncertainty is, since that uncertainty may be because a certain kind of ballot that has been counted extensively and another kind of ballot is not counted—by one kind, I mean mail-in versus in-person. And that may shift the numbers. And so, we’re going to try to be as open as possible to explain, “It looks like Candidate A here is in the lead, but we know that there’s a heck of a lot of ballots out there and from every piece of data and we know those ballots could easily tilt in the other direction.” So, that’s how we differentiate and how do we try to tell the story with as much transparency as possible.

SUAREZ: So often on election night, we’re dealing with the numerator and trying to get the numerator to be a bigger piece of the denominator. We won’t even know what the denominator is until pretty late in the game. Who was trying to jump in there?

BLANTON: Oh, it was Dana. I was just going to say that when it is one of these too close to call scenarios, that’s an excellent time, that’s when the poll team gets a lot of airtime because we can explain, “Well, here’s why it’s too close. Here’s why it’s tight.” Right? People, if they say the economy is the most important thing, maybe they’re going this way. If the virus is the most important thing, maybe they’re voting that, you know? And so, you have this, “Well, this group is going this way, and this group is going this way, and that’s why the race is so tight. And that’s why we don’t know, and we can’t tell you right now.” So that’s a good scenario for the poll team.

SUAREZ: David wants to know: “Your assumption is that the votes you’re using are, in fact, accurate. How do you verify that they are, and that the reporting of the vote is accurate?” This gets back to show your work and transparency. Where did the numbers come from in the first place?

SCOTT: They come from a lot of places. I’m sure the NDP is doing this too—at AP, the amount of effort that we put into doing quality assurance and quality checks is tremendous. At AP, our vote tabulation team—we are in thousands of local reporting units with people who are getting the vote totals directly from county election officials, directly from town clerks, and those are sent back to us. They do it by phone, right? Because one of the things that we learned in some testing is that you give someone a tablet and they type it in wrong and that’s harder to correct, and that’s harder to fix. And so, we do collect all that by phone. A lot of counties put their results up on the web, and so, we can grab information from there. One of my favorite stories is about a county in Kansas that just does a picture of a whiteboard posted to Facebook. Okay, that’s how you want to report your vote; we’ll find it and we’ll get it and we’ll put it into it, our system.

And there’s some other ways as well, but then we just do a tremendous amount of quality checks. There’s a hundred ballots, or a hundred votes being reported in this county. There’s only 80 people registered there. That’s a red flag. We need to go and look at that. This county is reporting a 60-40 for Trump, right? But the registration for that county is 90-10 for Democrats. That’s a red flag. We need to go back and look at that. And that type of constant quality check, and looking and double-looking at your work, and triple-looking at your work before you report out those numbers sort of gets back to something that Sam was saying earlier about the amount of investment that we put into it, the amount of resources we have on this, and the money that we spend. The elections research unit at AP is not something that we spin up once every four years. It’s a full-time job for a large number of people all of the time, so that we have a deep understanding of the rules in every county, the rules in every state, what their processes are. We have personal relationships with these town clerks and these county election officials that have been built up over many, many years. If you want to know the type of voting machines and Cuyahoga County in a particular precinct, the guys in AP elections research, they can tell you. That type of research, that type of prep work, when you match it up with all the quality checks that we do, gives me on the decision desk a really high level of confidence that what you’re seeing from our vote count team and in our vote count is the right number.

SUAREZ: Clara and Walter want to ask about keeping track of precinct reporting. “The percentage of precincts reporting seems like an anachronism in an era of mail balloting. Will you continue to show that graphic, especially when 99 percent could report, but a high percentage of absentee ballots remain uncounted?”

BUZBEE: I just want to say I am impressed by the sophisticated level of the questions, I just wanted to say you have a great group of questioners here.

SUAREZ: Go ahead, Arnon.

MISHKIN: The simple answer to the question is yes, we are not for any statewide race, we are just showing percent of expected vote, and by the way, it is an estimate I am making. There’s an over-under bet we have on how often during the night we’re going to have to increase the expected vote in the state because the turnout is going to be so high this year, possibly. So it’s an estimate, but that’s all we’re focused on, and we’ve redesigned our models around using percent of expected vote rather than percent of precincts reporting because yes, percent of precinct reporting is an anachronism in almost every jurisdiction in America.

FEIST: Ray, I think that’s across the board. I think that all of the major news organizations had switched to the percentage of expected vote. You see in maybe a green bar that grows across the screen on a vote board or online. And that’s the way it’s done because precinct reporting really is a bit of a thing of the past in terms of the way we view the vote coming in.

BUZBEE: And by the way, this is one of the things that we will try to aggressively and transparently talk about on election night, because this is something where things have changed and perhaps the public hasn’t quite caught up to that. So, we have a lot of explanatory journalism planned for election night around this very question.

SUAREZ: Patrick asks, “In the New York primary elections this year, Eliot Engle and Carolyn Maloney’s seats come to mind, it took weeks to call those races because of the slowness in counting mail-in and absentee. Might that happen again, and how do you plan for that should it happen on an even larger scale in a couple of states?” Good question.

AGIESTA: I can talk a little bit about that. New York is a little bit of an anomaly in how they count their votes. They were one of the only states in the entire country that does not count any absentee votes until after election night. So, what you’ll see in New York on November 3rd is going to represent Election Day vote, and they do not plan to report that absentee vote until later. Almost every other state in the entire country will give you some absentee and early voting on election night. It’s an extremely unique situation in New York. So, I wouldn’t use those primaries as an example of what we expect to see on November 3rd, everywhere.

SUAREZ: My mother was a poll lady for many years, and I used to make extra money by buying lunch for all the poll ladies and distributing bagged lunches to women who were chained to their desks on Election Day. And the absentee ballots didn’t start getting counted until all the machines were locked down and all the results from the machines were tabulated. Todd wants to ask about the post-election night period. He says, “We’ve all read the reports about how President Trump plans to use uncertainty, accusations of fraud to potentially prolong and complicate. The uncalled period after November 3rd. What’s your thinking about amplifying disinformation or democracy undermining information, even if it comes from the apex of power?  How are you thinking about how to report on it?”

BUZBEE: That’s another excellent question. And certainly something we’ve all been thinking about—and we have, obviously, regardless of what side of the political divide in the United States, any false claims or early claims you may come from. I think what we’re going to do is we’re going to stand by our journalism. We’re going to try to make our journalism as clear and accessible and explanatory as possible. We’re going to be very aggressive with that journalism. We’re going to say, “Here’s what the vote count shows—anyone who claims differently is not speaking accurately, according to the facts.” And we’re going to try to do that in real time. We’re in a somewhat unique position because our journalism goes to other news organizations, and then they present it. But on our mobile app and in the journalism we give to other news organizations, we’re going to give them the tools to combat misinformation in real time, regardless of where it comes from—any part of the known political spectrum in the United States. And that is one of the things that we will do. We do understand that a vacuum of information, because it does take some time to count the vote, is a potentially dangerous thing. And we have, as I said, accessible explanatory journalism that is prepared to say, “This is where we are in the count of the vote. We’re not here at the end, we’re here in the middle. And we will do our best. That’s all we can say, warning.”

SUAREZ: Arnon, go ahead.

MISHKIN: Well, I was just going to say Sally handled it very extremely well as only the executive editor of the AP really can, but I think that this goes to the need for transparency. And transparency means both that you don’t say what you don’t know, but also you say what you do know. And so, in the period after the election, we will have some sense of what we’re waiting on and the extent to which it points in one direction or another, or it points completely into their direction of uncertainty. And I think that the value that all news organizations can provide is in sharing that information to enable people to judge whatever is being said by whichever side as to whether or not they’re the level of truth that exists behind it.

SUAREZ: In real life, concession speeches, victory declarations, they’re not official acts. When somebody claims victory, that doesn’t mean they’re the winner. Just like, when someone concedes, that doesn’t mean they’re the loser, yet we covered them. Cameras go hot to election headquarters. When somebody walks up to a lectern—either a senior member of the campaign or the candidate herself or himself—we cover those events like events that have something to say about the result of the race. Well, we have to make a distinction in this race between what people are saying about it—to piggyback on Sally’s answer—and what we know to be true. Even if it goes as far as following a victory speech by saying, “We don’t know that this person has won” or following a concession speech by saying, “Live on the air. We don’t know that this person is lost.

FEIST: Right, without a doubt. And I think everybody will treat it the same way. If somebody were to clarify where to declare victory on election night before they had actually won—before the news organizations had projected winners—and they had accumulated 270 electoral votes, we will all note that the facts do not support this declaration. And we can explain why. We have so much data, so much information. Americans know how we elect presidents. You have to get to 270 electoral votes. That means you need a certain number of states. And if you don’t have those states, you haven’t yet won. And going back to something we touched on earlier, when election night ends—whether for you that’s two in the morning or four in the morning, or for us, sometime the next morning—the counting hasn’t ended. And so, if there’s not a winner in a particular state, the counting continues. And the people who voted by mail have just as much of a right to have their vote counted as the person who voted on Election Day.

And so, if a state is uncalled—because we don’t know the results of those other completely legitimate votes that will come in as the envelopes are opened, as they’re counted—we will wait, and we will let everybody know, as soon as we know. And so, this notion that some have suggested that we need to know the winner on election night—well, that’s just not true. None of us agrees with that notion. And we’ve seen two elections out of the last five where we didn’t know the winner on election night. So, if we know the winner on election night, terrific. If we don’t know the winner on election night, we will tell you as soon as we know, and it may be Wednesday. It may be Thursday. It may be a little bit later. But the votes that are counted the next day are equally important to the votes that were counted on Election Day.

SCOTT: If I could just add something to what Sam said about declarations of victory. The same rules apply to concessions, right. Candidates take back concessions, as we know, in some pretty high-profile races. So, just like you have to take care and do all the explanatory work around a declaration of victory, it’s the same thing with the concession. And so, in an AP story that might report about someone conceding a race, if we haven’t called it, that’s the next sentence—that AP has not called the race, and here are all the reasons why we haven’t.

MISHKIN: I think that we’ve described two kinds of speeches that candidates could give: victory speeches and concession speeches. There are third kinds. There’s a famous incident where Nixon made a mistake in 1960 and started talking about what he thought the election was looking like, sometime I think at eight in the evening California time; I was a little young. And there was a major mistake on his part, and ever since then, candidates have been known to be quiet until it’s over. I suspect that this year, there may be some examples of important statements made by not necessarily the candidate, but very high folks within a campaign, that could make life a little more complicated from a coverage perspective. And I think from a decision desk perspective, it’s exactly what David says, which is you focus on what the numbers show and what that means. And I think what Sam has said as well, and that’s what our focus is. But I think others of us, other folks within our organizations are going to have to think about how you handle all those statements that are coming out.

SUAREZ: I know several of you have asked, as Rachel has reiterated, “that a later result is not necessarily indicative of an issue with ballots or voting. And I recognize that this election will likely take longer. That being said, how would we know if there was some sort of issue, and how long can we reasonably expect this tallying to go on? Before it becomes something to worry about.” Good question, Rachel.

BUZBEE: I think that’s an excellent question. And I think there is, as Jenn pointed to, some uncertainty in this election, in the sense that there are states that have not handled large numbers of mail-in ballots in the past. And that is one of the things we’re all very focused on. One somewhat optimistic thing is that the states that were new to mail-in balloting seemed to get more good at it as time went on, a little bit. And that’s one of the things that we’ve seen, just a little bit. What I would say is that, first of all, one very important thing: It’s also possible that there would be a winner by a very large margin in this election. We don’t know. That would probably be a faster race call for the obvious reasons. It might not be. So, we don’t know for sure that it’s going to be lengthy, but we all suspect that there is a greater chance this year that it could be lengthy than in some of the past times we’ve seen.

I think that we will try to be as transparent as possible and as factual as possible about if there are some bigger problems, like, is a state just going slow counting mail-in ballots? Or, are they actually having a cybersecurity problem? We will be aggressively reporting and pushing really hard to find out that kind of information. I also think, and I don’t know, it might be interesting to hear from others on the call, but I think state election officials feel a pretty big burden this year, or responsibility this year, to be fairly transparent about what’s going on in their states. And hopefully they do do that, because I think that would benefit the American public, but our reporters will be trying to aggressively report if a state is going slow, is there a problem, or is it just that they’re counting mail-in ballots slowly? My suspicion is most of the time, it’ll just be mail-in ballots being counted slowly, but if there are problems, we promise we will transparently report that.

FEIST: And Ray, problems don’t necessarily mean something nefarious. We all dealt with the Iowa caucuses earlier this year, where the app that the state party used didn’t work. It was a mistake. It was a programming problem, but it was not a nefarious act. It wasn’t a nefarious actor in the end. They had all the information, it just took them a while to manually tabulate it and report it. So, problems don’t necessarily mean that somebody is trying to make mischief. It could just be something like what happened in Iowa, and you have a technical malfunction.

SUAREZ: Thanks for that reminder, Sam. Thank you to all the participants and our virtual lecture hall who’ve hung in there ’til the end and asked such terrific questions, and thanks to our panelists. Good luck on election night to all of you. And now, we’ll close with Suzanne Nossel from PEN. Suzanne?

NOSSEL: Just very quickly. I think we saw from the huge list of questions and the excellent questions what an engaging and important conversation this was. I’m extremely grateful to all of our participants for being here with us and working so hard to elucidate these critical aspects of this unprecedented election season, so thanks to everybody. We’re going to be posting this conversation on our website—it’ll be up on our YouTube channel. And we’re also going to record all the questions and to the extent that we can get answers for those that we didn’t get to today, we’ll make every effort to do that, hopefully working with the help of our panelists. So, thanks in advance for helping us fill in those holes and have a great afternoon, everyone.