This year will be my first CNN presidential election night. And in the lead up, I've been taking the assignment on the road. My last stop is Studio B in our Washington bureau. That's the home of CNN Politics. Welcome to the assignment table.

Do you take this all over?

I do take it all over. It's heavy, by the way. I wanted to be transparent with you about this because TV networks have a role to play on election night, or at least the way we've all experienced it the last few years.

Do you think it's funny that we rehearse election night?

I was actually afraid to say the word rehearse because I thought people would think that, you know, it was predetermined or something. So I'm sitting down with my friend David Chalian, CNN's political director. He's also been super kind to me as a television newbie. I mean, you tell me. I'm like, am I practicing sitting? How does it work?

It's well, because there are so many variables and there are so many components to it. It is the word choreography is correct. That's what we're rehearsing basically, is the choreography. But, you know, we get test data sent to us to help us rehearse election night. Every network does.

Chalian has seen a lot of nail biter elections. And from the vantage point, most of us haven't. The moment before the network decides to announce which candidate has won a district or a county or a swing state. News outlets like the Associated Press and CNN and Fox, well, they often make that call before a state has finished tallying the vote. And I wanted to know how? What goes into the decision making at the decision desk and what lessons have been learned from the election surprises of the last few years. I'm Audie Cornish. And this is the Assignment.

All right. So we're going to talk about just the whole idea of the decision desk, so to speak, because like every network has one. You and I both worked in multiple places, so we know it's a big operation wherever you go. And one of the things that's interesting is we take it for granted being in the business that the map gets populated, so to speak, fairly quickly, meaning how is it that any given network can look at a state and see that it has 10% of the vote in and then decide to like, call that state for a candidate?

So the only way in which you see projections with that little amount of voting is if there is not a competitive race. So all the prior polling prior to Election Day shows a blowout. I'm talking 30 points, 40 points, 50 points or it's uncontested. Literally. There's only one candidate on the ballot in the race of the two major parties. Those are what we call poll closing projections, meaning as soon as you start seeing the vote come in and you said, let's say 10%, if it is to form, if we see things that make us say this looks normal, that should be a Democratic county, it makes sense that there's Democratic vote coming in there, that everything about history tells us that's a deep Republican county. So the way that the vote just starts populating makes sense to everything we understand. And we start seeing margins that mirror these blowout polls. There is immediate confidence built into the system. But I just want to stress one thing at the outset here, Audie. The standard for making a projection remains the same no matter what, whether it's a blowout or it's the last state to decide the presidential contest, is that our decision desk, which is made up of statisticians and political demographic and geographic experts who understand this country and its politics and voting behavior. County by county, congressional district by congressional district. That team of people need to get to 99.9% certainty that a second place candidate is not does not have a mathematical possibility to overtake the first place.

So let's let's talk about that a little more and a little more specifically, because every network has its own secret sauce of coming to that certainty. And there's sometimes a mix of like exit polls, online surveys, your own polling that a network may have paid for to have done. I mean, you don't have to divulge too much, but like, what are the things that go into the decision here?

Exactly. So what you just said, exit poll information that comes in that our exit polls incorporate both talking to voters coming out of the voting booth on Election Day, but also incorporate a pre Election Day sort of telephone survey for populations that vote early basically so that. So in other words, the exit polls incorporate both early voters and Election Day voters. Right? So that information feeds in the vote, returns feed in. We have people that are reporting out vote counting in specific precincts that feed into our models. Right. That are sample precincts.

But the idea is you're kind of like bringing your own data to figure this out.

And real vote, though, and actual voting. So it's not just ours. Right. But so real vote that comes in in these sample precincts, the exit poll information, what we know about the polling in a race prior to Election Day, like what the state of the race was in the total aggregate of of polling out there, reliable polling that we use and what we know about the demography and geography of each of these places.

And so at this point, is that like you, some really gnarly looking cold cuts, like just in a room huddled over laptops? Is that printouts? Is that like I think people don't have a real understanding of what that how much of that decision is like spit out by a computer and its model and how much of it is a hive mind of people sitting in a room looking at it and then making a call?

I mean. They're sitting in an open newsroom. It's a team of a dozen or so folks that.

But I wouldn't come up to you in the middle of like.

I know and I'm not there. I'm on set for my role on election night, but I am in constant communication with them throughout the night. But I'm not sitting with them. But this is all here at CNN, headed up by Jennifer Agiesta, our director of Polling and Election Analytics, and she runs this team of decision desk analysts. And it is it is both things. It is, yes. A computer is constantly using the models that have been created to get to projection certainty, as you said. But of course, it is the the reason that it's populated with the human beings that have this expertise, whether it's expertise in House races or Senate races or presidential contests or certain regions of the country, they bring all of that to their statistical analysis. So it is both the real art and science.

So I heard from you and I've read from about some other decision desk folks that other networks there's sort of some questions people will have to ask themselves before they say this is the call. And one of them is, hey, can can candidate B actually catch up to. A reign based on everything we've seen. Can they catch up? Is there another 1 or 2 questions that you feel like kind of have to be answered or that people use as a litmus test to figure out, yes. It's time for us to say yes.

It really is that question that you just said. That is the ultimate question. But the questions that get asked prior to that question, that feed the answer to can candidate B overtake candidate A? Where is the vote? Still out. What is not yet counted in this race. And what do we know about that vote? What do we know about it? By method? Is it mail vote? Is it Election Day vote? That is still out. Is it overseas military ballots that are still out? What do we know in very close contests? Are there provisional ballots remaining to be counted? So they when we are down, like in 2020, when it's day after day and we're not able to project all the states yet to get a candidate at 270 electoral votes. At that point, we're it's a lot less about our models and much more about our reporting on the ground in these battleground states.

But the reason I ask because those outstanding questions are actually we realize they're more complicated than we thought in general. By we, I mean as an industry. So in 2020, you can't do a bunch of exit polls when everyone is doing mail in voting. Right. Because no one's walking to the polls. They've already sent them in. This was, I think, something that wasn't always reflected in the way pundits were talking on air, not ours. And it just in general about the race. The punditry and the polling are not the same thing, I guess, is what I'm asking. And so how do you kind of like manage the expectations people have about how something is going to turn out?

Well, first of all, you do it with humility. I have no idea how it's going to turn out. So this is the race we're covering now is a toss up race. I think it would be a fool's errand given how close each of these seven battleground states are. For someone to say that they know what the outcome of this race is going to be. I don't believe in my reporting with both campaigns that either campaign believes they know what the outcome of this race is going to be. And I know that, you know, right now the Trump campaign side and Republicans aligned with Trump are speaking with a lot more bravado. I don't believe that that is based in data necessarily.

And which is important to say. I think some people it's very campaigns have a vested interest in talking about the data they have in the best light. And so I think sometimes you hear their surrogates, you know, on TV, in newspapers, and it carries an air of, well, they must have some secret information.

Right now. I do think campaigns pay a ton of money for polling and data analytics, far more so than any news organization has to spend on this stuff. Right. And a lot of campaigns love to poo poo the public polls that we do because they feel they're spending a lot of money on that. But it is part of our reporting standards here at CNN. We don't report on internal polls of campaigns because we don't have insight into them. They're not transparent to us. I mean, for us to approve of a poll to report on, just like any other story we do, we have to vet the source material. Right. And we have to meet CNN's standards to go with it. And that is the same of a poll. So we do intense vetting. Our polling team does, of all the public polling out there available to us to report on and the ones that meet all of our benchmarks. And we are able to look inside and under the hood and see the methodology. And it meets our standards for reporting on a poll. Those are the polls we include in our poll of polls or that we would report on on a given day on our air or online. But we don't do that even though all of our sources are telling us what their internal polls say. We don't put those numbers on air because we can't verify.

Yeah. You know, it's funny, you mentioned 2020 and Dana Bash was quoted talking about 2020, saying it's like watching a mural being painted. And at the beginning it looks like one thing and then you realize it's something completely different from what you thought it was at the beginning. That's not normally how election nights are. How unusual was 2020? And just from your perspective, this just who's in this particular chair? What was unusual about it? Not what was perceived to be unusual? What was unusual?

I mean, what was unusual about 2020 was what you just said a few moments ago, which was a dramatic increase in mail voting in the way people were casting ballots.

Because of the Covid era, because.

We're in the midst of a pandemic and that there were a lot of states like Pennsylvania, for example, that didn't have a lot of practice with mail votes. And it took them a long time to process and count the ballots. And some of it is legislative driven, like in Pennsylvania. The law of the commonwealth is they cannot process open, even flatten the envelope that a ballot is in until Election Day. Now, in some other states, you can start processing, not counting the ballots, but you can start taking them out of the envelope, getting them in the right pile ready to feed into a machine. That doesn't happen until Election Day in Pennsylvania. So that to me, 2020 was teaching us, okay, these elections officials and God bless them for doing this work. They were encountering people casting their ballots in an entirely new way in great numbers. That is what took us so long to get results. I just want to say one of the things, though, about I think the larger point that in that Dana Bash quote that you just said. And this gets to your question about how do you sort of frame expectations? So not only do I have to be humble that in this race, we have no idea who's going to end up the winner here. But also, during the course of the evening of election night, as the returns are coming, we all sort of humility that the story forms, we get our guts as reporters is to constantly be like this piece of information came in so I can report this with authority. This takes time to unfold. Look at this map. Like these votes have to come in across the country. It takes time. It evolves. We have battleground states. They're out west. We have battleground states that closed late in the Great Lakes.

That instinct is our instincts. Like when I talked to the Arizona election officials, he's like, it's accuracy, it's transparency. Those are my founding forward values. And speed is a distant third place. And I feel like for media, that's not how we're trained. It's like you want to be the first to say, I.

Want to separate what I'm talking about.

Okay. You know.

Just to say I'm not talking about the speed of a projection right now, okay. Because the 2000 debacle of the networks getting Florida wrong multiple times, and then network presidents having to be hauled up to Congress to have to testify before Congress about why the network decision teams failed at their jobs, like nobody wants to, that's in this business wants to recreate that scenario again. So I think post 2000, speed has taken a far backseat on projections for the networks and for the Associated Press. I mean, that's part of also why we saw it take days in 2020. Nobody was looking to race ahead as all these ballots were still being counted. What I'm talking about is what is the narrative of the evening? I'm talking about the as the votes are coming in, and we're not in the zone of making a projection on these states yet, how are we forming the story of what we're seeing? And I think we have to have some patience for the voters to give, take a little time to tell us that story about what they're saying with their vote. That's all.

What's it like trying to get people to be patient?

Not easy.

Yeah. I mean, I can imagine. You mentioned 2000. Are there other times you can think back that were either tense or that the industry kind of had to reevaluate?

I mean, in 2018, in the midterm elections. Which was shaping up to be a Democratic year, we had to redo our midterm election night, basically, because one story took shape, that Democrats didn't do as well as they thought they were going to do. There were some high profile statewide candidates in places like Florida, Andrew Gillum running against DeSantis, or Stacey Abrams in Georgia, or Beto O'Rourke in Texas against Ted Cruz for the Senate. And those three high profile Democratic candidates,

Big narrative stories.

They were huge stories throughout that cycle. Democrats sort of performing better in these red states that, you know — and then it didn't come ashore that way. And so there was all this talk in real time of like, well, I guess this didn't turn out. And then California counted its votes and there was a huge Democratic takeover of the House of Representatives. Right. And, you know, Nancy Pelosi became Speaker again. So I just, this this stuff takes a little time to get the full clear picture. And so we went back in 2018 and did basically a second election night a week later when we had all the votes in to properly convey the story of what had occurred.

I'm talking with CNN's political director, David Chalian, about what to expect on election night. After the break, when are we going to know who the next president will be? Well, stay with us.

What's your election morning ritual?

That's a really good question. I don't have like, I work out at this time and I have this specific breakfast.

Everyone else is like just drinking caffeine. That's what they're telling me.

'Well, that for sure. And it is, you know, it's such a grind to get there. But our former bureau chief here at CNN, Sam Feist, who now is at C-Span, every election eve, would encourage us, this is now going back to the 2000 election, to read the internal CNN report about what went wrong, the internal investigation on election night 2000, and how the decision desk called the race incorrectly. And rereading that report on the eve of every election is very humbling, and it is part of the ritual of preparing for election day.

Dave, I have to ask, what are three take away points from that report?

Oh my God.

What can you, what are you allowed to say?

Everything I just described to you about what it takes for the decision desk to get to their level of certainty and all the expertise involved, there were just missed moments of certain bits of outstanding vote or there were just, there were, in the excavation looking backwards, whether it was a given county of information, somebody actually at one point in put the wrong numbers in a certain county. And that's, there are so many ways in which it went careening off course. So it's just a reminder about how we have to check, double check, quadruple check and get to that near 100% certainty mark before making a projection.

Fox famously called Arizona for Joe Biden in 2020, and there were a lot of claims that that was too early or wrong. The numbers turned out to be right. Joe Biden ended up winning that state by more than 10,000 votes.

Not just Fox. The Associated Press as well.

Thank you. Thank you. Yeah. And one of the former political editors who was involved in that call, he was let go from the company, but he's talked a lot now about the backlash from viewers. And there was just one thing I want to run by you. First, he describes them as being guys with a cool computer, lots of polling data and a lot of nicotine gum and coffee, which I just like the visual of that. But he also said, quote, Viewers had become even more accustomed to flattery and less willing to hear news that challenged their expectations. Me serving up green beans to viewers who had been spoon fed ice cream sundaes for years came as a terrible shock to their systems. And the reason why I wanted to ask you about this is because we talked about the punditry that happens in the months rolling up to election. The audience does build up certain expectations, right, based on what team they're rooting for. And then you show up on election night and you're like, there's no teams. There's just a number. How have you come to think about that as times have gotten more partisan?

I think you have to get rededicated to your process the analytics, the numbers, the math, because that's ultimately what this is. It's a math equation.

But in the era of like misinformation, is it a more nerve wracking job?

I am a believer. Maybe you think differently. The best way to. Oppose misinformation and disinformation is with actual information. I can't control bad actors wanting to inject bad information into the ether. What we can control is constantly informing our audiences wherever they're finding us. What is the current vote count? What do we know about what is left to count? What have we projected and why? And what haven't we projected and why? And we just constantly use all the facts and data we have to tell that story. That's what's in my control, right? I can't control — so I try to tune out a lot of that stuff. Of course, this is happening in a more polarized environment, but, but that doesn't impact what our core job is on election night in terms of being able to project the results.

So what keeps you up at night?

I mean, the idea of us making a projection that we have to pull back. And correct is definitely something that rattles through my brain at night. In this time leading up to the election. Our team is in, it is a remarkable team led by a superior journalist in Jenn Agiesta, who does all this. It it's just, you know, it's hard not to think of the huge ramifications in this environment, especially, of getting it wrong. And we won't. But that's what keeps me up at night.

How have you thought about. Scenarios like a candidate kind of calling it for themselves, saying it's looking like I've won and kind of going out and using the power of the microphone to make that case.

Well, we have some experience on this score. 2020 was a good run through for a lot of these things. And what is interesting, Audie, is that 2020 didn't catch us by surprise on this. We we were prepared for it because Donald Trump, four months before election night, 2020, was laying the groundwork publicly that he was not going to accept the results if they weren't to his liking. So we deployed a team of reporters, correspondents, producers to the battleground states. Solely on the reporting mission of the casting and counting of ballots. A far more robust team than we had done previously. Part of it was the pandemic, and there were a lot of people voting differently. And so we needed to be all over that, report that. But part of it was this suspicion that Donald Trump was trying to seed the earth with in the months leading up to the election. And we are doing the same again this time. We will be fortified with incredible reporting out of these battleground states, both on the process of the casting and counting of ballots, but also in sourcing and talking to the key elections officials, as you did in Arizona, in these states and in the critical counties, and making sure we understand everything of what has been counted and what is left to be counted. That reporting is going to guide us through these moments.

So we're going to be in this studio about a week from the taping of this conversation. Do you have any tips like granola bars or...

Lots of water while you are well hydrated.

Got it.

And I do think like power bars are good. Anything you can do to get some protein along the way. But we'll be living also on like terrible candy and sugar highs from moment to moment as our energy fades and we need to pick it back up. I just hope we have enough information because we're not the only ones where we have to hand off to another team of people who will so we can sleep a little bit. And I just hope we have enough information to know, okay, is now a good time to like, this team can get some rest and this team can.

You're asking about sleep in a very, in a way that won't get us an OSHA violation. You're asking whether or not you're going to sleep. I think the answer is no.

Don't you need some sleep?

Do you think we will have the outcome of the 2024 election on election night or the day after?

I think as you and I sit here eight days out, given how close these battleground states are, I would add, given that Arizona and Nevada and Pennsylvania are likely to take some time to count all of their votes, I would say Tuesday night is not likely, but I don't rule out a Wednesday resolution.

I love you for answering that. When people ask me right now, I'm just like, who knows what time hath wrought. Like, I just, because I really feel like —

Well, but there are some knowable things.

Okay, tell me.

Like we know Milwaukee and Detroit. Two huge cities in two big battleground states, are not probably going to report their vote until 3 or 4:00 in the morning on Wednesday.

Right.

So it's tough to make a projection in those states if you don't have Wayne County or Milwaukee County. Right. So there are certain things we do know. We know Arizona takes time to count its votes. We know Pennsylvania won't be able to count all the mail ballots on election night. That's what that's what gives me the ability to say. I don't think Tuesday night is likely. But I would be surprised, I would be surprised — and I'm I'm always looking to be surprised on election week,

Same.

I love it. That, to me is like the joy of covering it. I don't ever want to be in a place where I think I know all the answers. I don't. The joy of covering this stuff is in the surprise. But I would be surprised if this goes all the way to Saturday like it did four years ago.

Well, David Chalian, thank you for letting me take some of your time when I know you're supposed to be somewhere looking at data. This is really means a lot. And it's actually super helpful going into the next couple of days.

Thanks for having me.

David Chalian, host of the podcast CNN Political Briefing. And he is political director here at CNN. The Assignment is a production of CNN Audio. If you liked it, we want you to hit the follow button. And it's super important that if you loved it, that you share it. We want more people to discover the show. This episode was produced by Lori Galarreta, Graelyn Brashear, Dan Bloom, and Osman Noor. The senior producer of the assignment is Matt Martinez. Our video team this week, John Anglim, Harlan Schmidt, Andrew Christman, and our video editor is Cole Deines. Dan Dzula is our technical director. Steve Lickteig is the executive producer of CNN Audio. And we had support from Haley Thomas, Alex Manasseri, Robert Mathers, John Dianora, Leni Steinhart, Jamus Andrest, Nicole Pesaru and Lisa Namerow. Special thanks, as always, to Katie Hinman. And thank you for listening. I'm Audie Cornish.