He Oversees Fox News’ Decision Desk. Here’s His Outlook for Election Night.
In a recent interview, Arnon Mishkin discusses the contentious election calls he has made and shares his forecasts for the future.
TroibNews 
NEW YORK — The individual who might announce the victor of the 2024 presidential election is prepared to give a timeline for when the announcement could occur.
“The over/under is Saturday,” stated Arnon Mishkin, the head of Fox News’ decision desk. “Which was when the call was made last time.”
In a recent interview with PMG Magazine at the News Corp. building in Midtown Manhattan, Mishkin expressed his intention to call the election as early as possible, though he acknowledged that a drawn-out vote count could extend the timeline to several days.
Regardless, he emphasized that he feels no pressure from the network's executives to “make our audience happy.”
At 69, Mishkin serves as a consultant and is not a direct employee of Fox, but he has led the network's decision desk for presidential elections since 2008 and manages a team of nine analysts.
While largely operating behind the scenes, Mishkin has been pivotal in several key moments during Fox’s election night broadcasts. Notably, in 2012, he engaged in an on-air debate with Karl Rove, the former political strategist for George W. Bush, when Rove challenged the network's projection that Barack Obama had won Ohio.
In 2020, he gained attention for defending Fox’s call of Arizona for Joe Biden — a decision some have criticized as premature, despite Biden winning the state by a narrow margin of just over 10,000 votes. This projection, along with Trump's attempts to reverse his defeat, stirred considerable controversy at Fox News, resulting in some staff departures.
Mishkin reaffirmed his commitment to the role, sharing insights about his typical Election Day routine and his family's reactions to a "Succession" episode featuring a character reminiscent of him.
This conversation has been condensed for clarity.
**When it comes down to calling a state, what are the inputs that go into that decision?**
There are two-and-a-half inputs. Half of the input is what the prior reporting says about the state, what the pre-election polling says, etc. The two main inputs that go into our models are No. 1 what we call the [Fox News Voter Analysis], which is the [voter] poll we conduct across all 50 states with the National Opinion Research Center. Because of the increase in different types of modalities in voting, we wanted one system that covered all types of voters: early voters, Election Day voters, and mail-in voters. And that’s the system that is known as VoteCast or the Fox News Voter Analysis.
The [second] thing you’re looking at is the actual vote. Over the course of the evening, the FNVA becomes very unimportant and basically what you’re looking at is the actual vote count.
And there’s a set of models that we use. I think only about 13 states report vote by vote type, but even if you don’t know vote type, you kind of know the way the state is reported. The bulk of the early vote count is going to be mail-in, or the bulk of the early vote count is going to be Election Day. North Carolina, it’s all mail-in. Pennsylvania, it’s all Election Day, primarily.
**Do you have a better understanding of how these states are going to report and how voters are going to participate going into this election than going into the last presidential election during the height of Covid?**
I think we had a pretty good understanding of how they were going to report [in 2020]. What we don’t know now — but we kind of knew then — was that there was going to be an enormous skew in the vote type. I believe that Joe Biden won every state in the mail-in vote with the exception of Alabama — and that Trump won every state on Election Day with the exception of Vermont and maybe Connecticut. I can’t remember exactly, but there was a huge skew. I think that skew is going to be less this time. I don’t know how much less. I know it’s still going to be there, but the Republicans are doing a much better job of encouraging people to vote however they want.
**So would you say that makes the job harder in 2024 than it was in 2020? Because we don’t necessarily know that, "Oh, all the Democrats are going to vote early or by mail, and all the Republicans are going to vote on Election Day."**
Yeah, I think it’s going to be a little harder. On the other hand, I think that some of our models have gotten better. We’ve created a new model based on the vote count which is much more focused on either getting vote by type and then using that in the model, or knowing that you’re not going to get vote by type and then making estimates around on that. And then I think our FNVA — because we ask people how they vote, it’s self-reported — we have a pretty good idea of what the skew is. And that number has been pretty accurate over the years.
**When do you think you’re going to be able to call this election? Do you think it’s going to be the night of? A week later? Possibly longer? The race seems so close.**
The race seems very, very close. It is dependent on a number of states, like Pennsylvania, that we believe are going to be reporting in a pattern similar to the way they have reported in the past. So I’d say, the over/under is Saturday. Which was when the call was made last time. Which is when Pennsylvania is likely to come in.
I think we have to accept the reality that we don’t really know how close this election is going to be. I’m pretty sure it’s going to be close. I see some polls that say, “Actually, it ain’t going to be close. It’s going to be one way or the other.” There’s some reporting that Trump is sort of gaining. Some of the polls have showed he’s gaining. There’s another sense I have that actually he may be declining. I think the real issue is what happens to Trump. I’ve always thought this about this election: It’s less about who’s running against him than it’s about Trump.
**But it’s his vote share, you mean? Which was 46 percent and 47 in his two elections.**
It was 46.1 percent in 2016. It was 46.7 or 46.8 in 2020.
46.1 and a very low turnout — his advantage in the Electoral College allowed him to win in 2016. With a very heavy turnout and him at roughly 47 in 2020, his Electoral College advantage meant it was a really close election in the Electoral College.
**Let’s talk about what comes after Election Day, and whether that affects when you decide one candidate has won a state. Given what we saw in 2020 from Trump with this really expansive litigation, efforts to designate alternate slates of electors, obviously the pressure to have Congress not accept the electoral votes on January 6, does that go into your thought process or affect how you do your job at all?**
No. Our job on the decision team is to say who is getting more votes. Is it Candidate X or Candidate Y? That’s our job.
If there are issues about the outstanding vote and the like, we take that into account. The litigation which exists is a news story. But the story we’re focused on is, who is getting more votes. The litigation may affect stuff. There’s already the start of litigation. I think all the journalistic organizations are focused on that. To the extent the election’s going to be decided if the courts make certain decisions, that’s a story. But it’s a different story from: Who got more votes?
**I want to ask you a little bit about 2020, and especially Arizona. Fox called the state for Biden before any other network or the Associated Press, which followed you a couple of hours later. That call ultimately proved correct. But a lot of analysts who’ve looked at it say that, in retrospect, the call came too soon, especially the way that some of those late votes broke. Do you think it was too early to have made the call when you all did?**
Again, our job is to say who’s going to win. And I’ve gone through this a couple of times. The models we use — there were two models we used. The most important model was the integrated model, which combined the poll data with the election reporting from the state. We knew that the bulk of the reporting from Arizona was mail-in at that stage. When the model compared the vote in [Arizona] to our poll data, it showed we were off by a point on the mail-in vote. We were comparing our estimate of the mail-in in the poll, and the actual, I think it was about 85 percent of the mail-in was in. And it was dead-on. And the integrated model combined the two and had an estimate.
It was clear that we had evidence that our poll was accurate in Arizona — maybe not in other states, but in Arizona we had evidence of that. And based on that, we estimated that Candidate X — in this case, Biden — had a lead that was four standard deviations from being able to be taken over. We had another model as well, the [outstanding vote needed] model, which said it was not going to be possible for Candidate B, who was in this case Donald Trump, to take over. He obviously came really, really close. But the fact that he didn’t overtake him — this is statistics. I think that the model proved to be right because he didn’t take over. So our minds were comfortable with that call. In the end, the stomachs were churning. But the mind still said [Biden] was going to win.
**The experience of watching Trump win about roughly 60 percent of the outstanding vote at that point — why did that happen? What does it tell you about the way Arizona reported its votes? And what can you learn about that going into 2024?**
I think what the story really tells is that Donald Trump is a massive get-out-the-vote machine. Whether it’s organized or not, he has a base of support, and they are ready to come out for him. The evidence from the 2018 and 2022 elections is that when he’s in the ether, he’s also a get-out-the-vote machine for the Democrats. And when he’s on the ballot, he’s a get-out-the-vote machine for both parties, but particularly for the Republicans. And I think that’s what we learned from Arizona. What basically happened was the poll was off in terms of what the Election Day vote or the Election Day submissions were going to be.
**You did an interview four years ago around this time with Ben Smith, who was then at The New York Times. And the print headline they ran it under was, “Will Fox News’s Chief Nerd Call It Straight on Nov. 3?” First of all, do you embrace the title of “Chief Nerd?”**
It’s tattooed. And my kids would say it’s accurate.
**But you also don’t see those headlines about your five counterparts at the other networks. There’s no, “Will NBC News’ John Lapinski Call It Straight?” Or “Will the AP’s Stephen Ohlemacher Call It Straight?” Why do you think that is?**
My charm.
**Do you feel more pressure given that Fox’s viewership leans more Republican than some of the other news organizations that are going to be part of this process of calling the election?**
No, I don’t feel any pressure because of that. I think the pressure the team feels is: Focus on the numbers, and focus on who’s going to win, and let us know. And I don’t feel any sort of pressure like, “Call it our way. Make our audience happy.”
**How about another form of pressure: Do you feel any pressure to be first? And right, but first? In my experience, Fox News and NBC are typically the earliest networks to call a lot of these races. Some of the other broadcast and cable networks seem to be a little more conservative. Is that a guiding principle here — to want to be the first to tell the country who the next president’s going to be?**
I don’t think so. I fessed up to the title of “nerd,” and the team is a bunch of college professor-types, statistics or political science. So smart people like to get the answer first, that’s just the way they are. But I don’t feel that that’s a pressure. I think that there’s a distinct pressure to be right — or, rather, not be wrong. So I think that’s the pressure. And yes, there have been some calls that have been faster than others — obviously — but at the same time I don’t feel any sort of pressure from that.
The public has a right to know who the next president is. Uncertainty about that is not a good thing. And anything you can do to contribute to the information about that is an important thing. And it would’ve been appropriate for us to share the information about that on election night in 2020.
And by the way: When we made that Arizona call, I did not think the election was over. I know some people thought the election was over. I thought that the election was still a dogfight in the Electoral College. Trump’s lead in Pennsylvania was something that — based on our data — it wasn’t clear that Biden could take over.
**I’ve read that you have an Election Day routine: You play the cello in the morning?**
I do. I’ve been bad about it, but I do have that routine. As one of my kids said, “Everyone else has a mid-life crisis and gets a motorcycle, and you buy yourself a cello.”
**So you took up the cello as an adult?**
I played piano and cello as a kid, but I re-took it, like, at age 50.
**And how did this become an election morning routine? I have one: I get a haircut on Election Day.**
It started because I got OK [at the cello], and I said, “I have to do that on Election Day.” Bach’s “Cello Suite No. 1.”
**Does it do something for your mindset?**
Oh, it takes your mind off what’s really going to happen, the anxiety and all the rest. It is cool.
**Last year, the HBO show “Succession” aired an episode about a presidential election night at a fictional cable news network that some people have compared to Fox News. Have you seen that episode?**
No.
**There’s a character in the episode who is the head of the decision desk at that network. Brian Stelter, the CNN media columnist, called the character “clearly a riff” on you, and that if he were you “I would probably have a very hard time watching this episode.” Have people told you that you should watch it, that you shouldn’t watch it?**
I can show you the family chat, and one of the kids did watch it and said, “He’s better looking than you. He’s much thinner.”
**That’s a good answer. No interest in watching it, though?**
No.
**2020 obviously did have a lot of controversy, and some of the internal conversations at Fox News about the Arizona call and the election ended up getting published in court filings. And there was a lot of turnover here at the decision desk. Did you think about not coming back for 2022 and 2024? Was that ever a consideration?**
Not really. I mean, you want to go back, and it is great fun. I enjoy it. It is very tension-provoking. But I guess I’m a stress junkie. So I did not think of that. There was a lot of controversy about the stuff that happened afterward. Whether we’d called Arizona or not, I don’t think that would’ve changed any of that at all.
Allen M Lee contributed to this report for TROIB News
“The over/under is Saturday,” stated Arnon Mishkin, the head of Fox News’ decision desk. “Which was when the call was made last time.”
In a recent interview with PMG Magazine at the News Corp. building in Midtown Manhattan, Mishkin expressed his intention to call the election as early as possible, though he acknowledged that a drawn-out vote count could extend the timeline to several days.
Regardless, he emphasized that he feels no pressure from the network's executives to “make our audience happy.”
At 69, Mishkin serves as a consultant and is not a direct employee of Fox, but he has led the network's decision desk for presidential elections since 2008 and manages a team of nine analysts.
While largely operating behind the scenes, Mishkin has been pivotal in several key moments during Fox’s election night broadcasts. Notably, in 2012, he engaged in an on-air debate with Karl Rove, the former political strategist for George W. Bush, when Rove challenged the network's projection that Barack Obama had won Ohio.
In 2020, he gained attention for defending Fox’s call of Arizona for Joe Biden — a decision some have criticized as premature, despite Biden winning the state by a narrow margin of just over 10,000 votes. This projection, along with Trump's attempts to reverse his defeat, stirred considerable controversy at Fox News, resulting in some staff departures.
Mishkin reaffirmed his commitment to the role, sharing insights about his typical Election Day routine and his family's reactions to a "Succession" episode featuring a character reminiscent of him.
This conversation has been condensed for clarity.
**When it comes down to calling a state, what are the inputs that go into that decision?**
There are two-and-a-half inputs. Half of the input is what the prior reporting says about the state, what the pre-election polling says, etc. The two main inputs that go into our models are No. 1 what we call the [Fox News Voter Analysis], which is the [voter] poll we conduct across all 50 states with the National Opinion Research Center. Because of the increase in different types of modalities in voting, we wanted one system that covered all types of voters: early voters, Election Day voters, and mail-in voters. And that’s the system that is known as VoteCast or the Fox News Voter Analysis.
The [second] thing you’re looking at is the actual vote. Over the course of the evening, the FNVA becomes very unimportant and basically what you’re looking at is the actual vote count.
And there’s a set of models that we use. I think only about 13 states report vote by vote type, but even if you don’t know vote type, you kind of know the way the state is reported. The bulk of the early vote count is going to be mail-in, or the bulk of the early vote count is going to be Election Day. North Carolina, it’s all mail-in. Pennsylvania, it’s all Election Day, primarily.
**Do you have a better understanding of how these states are going to report and how voters are going to participate going into this election than going into the last presidential election during the height of Covid?**
I think we had a pretty good understanding of how they were going to report [in 2020]. What we don’t know now — but we kind of knew then — was that there was going to be an enormous skew in the vote type. I believe that Joe Biden won every state in the mail-in vote with the exception of Alabama — and that Trump won every state on Election Day with the exception of Vermont and maybe Connecticut. I can’t remember exactly, but there was a huge skew. I think that skew is going to be less this time. I don’t know how much less. I know it’s still going to be there, but the Republicans are doing a much better job of encouraging people to vote however they want.
**So would you say that makes the job harder in 2024 than it was in 2020? Because we don’t necessarily know that, "Oh, all the Democrats are going to vote early or by mail, and all the Republicans are going to vote on Election Day."**
Yeah, I think it’s going to be a little harder. On the other hand, I think that some of our models have gotten better. We’ve created a new model based on the vote count which is much more focused on either getting vote by type and then using that in the model, or knowing that you’re not going to get vote by type and then making estimates around on that. And then I think our FNVA — because we ask people how they vote, it’s self-reported — we have a pretty good idea of what the skew is. And that number has been pretty accurate over the years.
**When do you think you’re going to be able to call this election? Do you think it’s going to be the night of? A week later? Possibly longer? The race seems so close.**
The race seems very, very close. It is dependent on a number of states, like Pennsylvania, that we believe are going to be reporting in a pattern similar to the way they have reported in the past. So I’d say, the over/under is Saturday. Which was when the call was made last time. Which is when Pennsylvania is likely to come in.
I think we have to accept the reality that we don’t really know how close this election is going to be. I’m pretty sure it’s going to be close. I see some polls that say, “Actually, it ain’t going to be close. It’s going to be one way or the other.” There’s some reporting that Trump is sort of gaining. Some of the polls have showed he’s gaining. There’s another sense I have that actually he may be declining. I think the real issue is what happens to Trump. I’ve always thought this about this election: It’s less about who’s running against him than it’s about Trump.
**But it’s his vote share, you mean? Which was 46 percent and 47 in his two elections.**
It was 46.1 percent in 2016. It was 46.7 or 46.8 in 2020.
46.1 and a very low turnout — his advantage in the Electoral College allowed him to win in 2016. With a very heavy turnout and him at roughly 47 in 2020, his Electoral College advantage meant it was a really close election in the Electoral College.
**Let’s talk about what comes after Election Day, and whether that affects when you decide one candidate has won a state. Given what we saw in 2020 from Trump with this really expansive litigation, efforts to designate alternate slates of electors, obviously the pressure to have Congress not accept the electoral votes on January 6, does that go into your thought process or affect how you do your job at all?**
No. Our job on the decision team is to say who is getting more votes. Is it Candidate X or Candidate Y? That’s our job.
If there are issues about the outstanding vote and the like, we take that into account. The litigation which exists is a news story. But the story we’re focused on is, who is getting more votes. The litigation may affect stuff. There’s already the start of litigation. I think all the journalistic organizations are focused on that. To the extent the election’s going to be decided if the courts make certain decisions, that’s a story. But it’s a different story from: Who got more votes?
**I want to ask you a little bit about 2020, and especially Arizona. Fox called the state for Biden before any other network or the Associated Press, which followed you a couple of hours later. That call ultimately proved correct. But a lot of analysts who’ve looked at it say that, in retrospect, the call came too soon, especially the way that some of those late votes broke. Do you think it was too early to have made the call when you all did?**
Again, our job is to say who’s going to win. And I’ve gone through this a couple of times. The models we use — there were two models we used. The most important model was the integrated model, which combined the poll data with the election reporting from the state. We knew that the bulk of the reporting from Arizona was mail-in at that stage. When the model compared the vote in [Arizona] to our poll data, it showed we were off by a point on the mail-in vote. We were comparing our estimate of the mail-in in the poll, and the actual, I think it was about 85 percent of the mail-in was in. And it was dead-on. And the integrated model combined the two and had an estimate.
It was clear that we had evidence that our poll was accurate in Arizona — maybe not in other states, but in Arizona we had evidence of that. And based on that, we estimated that Candidate X — in this case, Biden — had a lead that was four standard deviations from being able to be taken over. We had another model as well, the [outstanding vote needed] model, which said it was not going to be possible for Candidate B, who was in this case Donald Trump, to take over. He obviously came really, really close. But the fact that he didn’t overtake him — this is statistics. I think that the model proved to be right because he didn’t take over. So our minds were comfortable with that call. In the end, the stomachs were churning. But the mind still said [Biden] was going to win.
**The experience of watching Trump win about roughly 60 percent of the outstanding vote at that point — why did that happen? What does it tell you about the way Arizona reported its votes? And what can you learn about that going into 2024?**
I think what the story really tells is that Donald Trump is a massive get-out-the-vote machine. Whether it’s organized or not, he has a base of support, and they are ready to come out for him. The evidence from the 2018 and 2022 elections is that when he’s in the ether, he’s also a get-out-the-vote machine for the Democrats. And when he’s on the ballot, he’s a get-out-the-vote machine for both parties, but particularly for the Republicans. And I think that’s what we learned from Arizona. What basically happened was the poll was off in terms of what the Election Day vote or the Election Day submissions were going to be.
**You did an interview four years ago around this time with Ben Smith, who was then at The New York Times. And the print headline they ran it under was, “Will Fox News’s Chief Nerd Call It Straight on Nov. 3?” First of all, do you embrace the title of “Chief Nerd?”**
It’s tattooed. And my kids would say it’s accurate.
**But you also don’t see those headlines about your five counterparts at the other networks. There’s no, “Will NBC News’ John Lapinski Call It Straight?” Or “Will the AP’s Stephen Ohlemacher Call It Straight?” Why do you think that is?**
My charm.
**Do you feel more pressure given that Fox’s viewership leans more Republican than some of the other news organizations that are going to be part of this process of calling the election?**
No, I don’t feel any pressure because of that. I think the pressure the team feels is: Focus on the numbers, and focus on who’s going to win, and let us know. And I don’t feel any sort of pressure like, “Call it our way. Make our audience happy.”
**How about another form of pressure: Do you feel any pressure to be first? And right, but first? In my experience, Fox News and NBC are typically the earliest networks to call a lot of these races. Some of the other broadcast and cable networks seem to be a little more conservative. Is that a guiding principle here — to want to be the first to tell the country who the next president’s going to be?**
I don’t think so. I fessed up to the title of “nerd,” and the team is a bunch of college professor-types, statistics or political science. So smart people like to get the answer first, that’s just the way they are. But I don’t feel that that’s a pressure. I think that there’s a distinct pressure to be right — or, rather, not be wrong. So I think that’s the pressure. And yes, there have been some calls that have been faster than others — obviously — but at the same time I don’t feel any sort of pressure from that.
The public has a right to know who the next president is. Uncertainty about that is not a good thing. And anything you can do to contribute to the information about that is an important thing. And it would’ve been appropriate for us to share the information about that on election night in 2020.
And by the way: When we made that Arizona call, I did not think the election was over. I know some people thought the election was over. I thought that the election was still a dogfight in the Electoral College. Trump’s lead in Pennsylvania was something that — based on our data — it wasn’t clear that Biden could take over.
**I’ve read that you have an Election Day routine: You play the cello in the morning?**
I do. I’ve been bad about it, but I do have that routine. As one of my kids said, “Everyone else has a mid-life crisis and gets a motorcycle, and you buy yourself a cello.”
**So you took up the cello as an adult?**
I played piano and cello as a kid, but I re-took it, like, at age 50.
**And how did this become an election morning routine? I have one: I get a haircut on Election Day.**
It started because I got OK [at the cello], and I said, “I have to do that on Election Day.” Bach’s “Cello Suite No. 1.”
**Does it do something for your mindset?**
Oh, it takes your mind off what’s really going to happen, the anxiety and all the rest. It is cool.
**Last year, the HBO show “Succession” aired an episode about a presidential election night at a fictional cable news network that some people have compared to Fox News. Have you seen that episode?**
No.
**There’s a character in the episode who is the head of the decision desk at that network. Brian Stelter, the CNN media columnist, called the character “clearly a riff” on you, and that if he were you “I would probably have a very hard time watching this episode.” Have people told you that you should watch it, that you shouldn’t watch it?**
I can show you the family chat, and one of the kids did watch it and said, “He’s better looking than you. He’s much thinner.”
**That’s a good answer. No interest in watching it, though?**
No.
**2020 obviously did have a lot of controversy, and some of the internal conversations at Fox News about the Arizona call and the election ended up getting published in court filings. And there was a lot of turnover here at the decision desk. Did you think about not coming back for 2022 and 2024? Was that ever a consideration?**
Not really. I mean, you want to go back, and it is great fun. I enjoy it. It is very tension-provoking. But I guess I’m a stress junkie. So I did not think of that. There was a lot of controversy about the stuff that happened afterward. Whether we’d called Arizona or not, I don’t think that would’ve changed any of that at all.
Allen M Lee contributed to this report for TROIB News
