Arnon Mishkin is the director of the Fox News election night decision desk, a position he has held since 2008. He analyzes polls, election models, and real-time election results to call congressional and presidential races. Following the 2016 election, Mishkin worked with the Associated Press and the National Opinion Research Center to develop AP Vote Cast, an alternative to traditional exit polls for measuring public opinion on election day. He led the Fox decision desk team during the 2020 election, which controversially called Arizona for Joe Biden on election night. Previously, Mishkin worked under political consultant David Garth, contributed to election night coverage for NBC News, and served as a strategic consultant at the Boston Consulting Group and Mitchell Madison.
This interview, conducted by Ethan Jasny, has been edited for length and clarity.
Harvard Political Review: Fox News has arguably played a substantial role in the process of polarization. Your role is totally independent from the opinion and editorial side of the network, but from your perspective within Fox, how has the network’s polarizing content developed over time? And what do you think this tells us about polarization in America more broadly?
Arnon Mishkin: First of all, I’m on the news side of Fox, not the opinion side. And I’m a contractor of Fox, not a decision-maker of that sort. So, I know that there’s a lot of writing that says that basically Fox caused polarization, but I’m not certain that’s correct. We had Steve Kornacki from MSNBC at the JFK Jr. Forum. And he wrote a whole book, which is basically about the origins of polarization, stating that it really started in the 80s and the 90s with Newt Gingrich and the election of a Republican majority in the House. From 1954 to 1994, the Republicans were completely shut out of control of the House. In an environment where one party is shut out of being in the majority, the only strategy available to that party is: “If we’re going to have any influence on policy where we want to have influence, the only way we can do it is to basically compromise with the majority party because we’ll never get control of the majority.”
The concept in [Kornacki’s] book is that Newt Gingrich introduced a totally different way of operating as a minority from the days of [former Republican minority leader] Bob Michael. That was one of the reasons the Republicans believe that they wound up in the majority. [They came to believe that] when you’re fighting over majority control, the right strategy for either party is to be on the offensive and to say: “We disagree with what these guys are doing, and therefore we are pushing in one way or the other.” So I think that’s a major cause of the polarization. And by the way, all those steps we just mentioned predate the launch of Fox News in 1996.
As for the role of the media in polarization, I think it’s not so much that the media decided, or any agent of the media decided: “Let’s do this.” The whole nature of the media has changed. In the 70s and the 80s, it was a world of three networks. CNN had started in 1980, but it was basically irrelevant until 1991 from a ratings perspective. Walter Cronkite used to get 30 million viewers a night. I believe that Fox broke the record for election night coverage in 2020, and we had 14 million. In an environment where you’re reaching 30 million a night, your inherent need is to try to offer as broad a perspective as possible. In an environment where you’re reaching 1 or 2 million a night (and this goes for a Fox, MSNBC, and CNN), your motivation is to pursue an audience by providing a very specific perspective.
It’s not just true of the television media; it’s also true of the print (or the formerly print) media. So the New York Times, the Washington Post, and all these newspapers used to just be as broad an audience as possible in order to get advertising revenue. Now it’s: “We need an audience that’s willing to pay for our content, and pay hundreds of dollars a year for the content.” As the newspapers have moved to the subscription business model, they have moved to pursuing a perspective, because that’s what people pay for: “I want to read stuff I agree with.”
HPR: This is the chicken-or-the-egg question of polarization. What happens first: the polarization of ‘regular citizens’ or the polarization of elites in the media? Obviously it’s a dynamic, ongoing process. But would you say it’s more of a top-down or bottom-up process?
Mishkin: My own belief, and this is not meant to sound cynical, is that politicians are intellectual entrepreneurs. There are two ways to lead a parade. One is to hold up a flag and gather an audience behind you, and say: “Let’s march.” And another is to see where people are marching, and run to the head of the line, and say: “I’m leading this parade.” And I think that politics is a mixture of those two things. If you’ve talked to the folks in the government department here, there is a whole concept that the role of a politician is: “I have this group, I need to represent their views right, as closely as possible.” But if you push me to give me a number, the most effective politicians are 90% “Where’s the parade going? I want to get into the head of it,” and 10% ‘Okay, let me try to move them slightly in this direction.”
HPR: One word that gets tossed around a lot is tribalism. For example, in Ezra Klein’s book Why We’re Polarized, he goes through the behavioral psychology literature and makes the argument that humans have a natural tendency to form in- and out-groups based on identity. Do you find this almost evolutionary conception of polarization convincing?
Mishkin: I think it’s a really interesting concept. And it dovetails with another theory of mine. If you date polarization as Kornacki does, basically 1992 was the first presidential election in two generations that was fought without the specter of the Soviet Union. Ever since World War Two, there was a view that the U.S. was head-to-head against the Soviet Union: that is the enemy. And we need a president who is capable of standing up to that enemy. That’s what all these elections were fought over: who do you want in charge of the nuclear codes? In 1990, the Soviet Union collapsed; the Cold War has ended. Lo and behold, the next presidential election, we pick up this governor from Arkansas, who’s had no experience in foreign policy, who has all sorts of iffy things. But on the other hand, he’s going to show you an interesting time.
And ever since then, with the possible exception of 2004 and 9/11, elections have not been fought with the idea of a major foreign adversary. And so why did these two points dovetail: tribalism and the Cold War? Throughout the Cold War, there was tribalism regardless of who you were in America — whether you were some college-educated elite in Cambridge, Massachusetts or a kid in a very rural area — you were united by the idea that you’re an American, and that you’re fighting the Soviet Union. And I think that with the absence of that foreign enemy, it’s much easier to form sub-tribes.
HPR: Today we see a partisan bifurcation of people’s media sources: for example, conservatives are more likely to get their news from sources like Fox, while liberals are more likely to get their news from sources like CNN or MSNBC. But do you think there is also a bifurcation of the way that media is spread and packaged? Do you think there are differences in the forms of media and the way that media is consumed for conservatives and liberals?
Mishkin: For the past 40 years, talk radio has been a very important Republican vehicle of communication. The internet is an increasingly Democratic or more liberal form. It gets more affluent, college-educated types. Although, some conservatives do very well with Facebook links. But I don’t know if it’s so much form. I think that one of the real problems we see is that there’s this asymmetric story coverage. There’s an asymmetric politics that is part of the source of polarization. And the asymmetry is that Republicans are fundamentally focused on a very different set of issues than Democrats. When you look at the coverage of what’s going on today, you see that media that attracts more of a Republican audience focuses on very different story selection than media that attracts the Democratic audience.
I think it would be useful for everyone to understand what those stories are. You hear in more conservative media about things like the border, taxes, crime, inflation. And on the Democratic side, all they’re talking about is what’s going on with the reconciliation and infrastructure bill. I think it would be useful in both cases if they looked at the stories that are of interest to the other side, if for no better reason than to understand why they’re important. Because right now, when you hear the Democrats talking about what’s going on at the border, they just say, “the Republicans are trying to distract attention.” But maybe they’re not — maybe [the border is] an issue for a lot of people. And you should focus on why that’s an issue or why it isn’t an issue, but you shouldn’t just ignore it because then we have what happened in 2016, when all of a sudden Donald Trump is elected [while] a lot of media outlets were assuring the audience, “Don’t worry, there’s no circumstance under which Hillary Clinton can lose.”
HPR: Polling in 2020 — and not just for the presidential race — seemed to skew Democratic for the second cycle in a row in terms of undervaluing Trump’s chances. And even as many firms in 2020 adopted practices such as weighting by education, which was thought to be a major methodological flaw in 2016, the polls still seem to underestimate Trump. Why do you think this is? Do you think that this is a Trump-related issue mainly? Could it play out differently in the upcoming midterms?
Mishkin: There’s been an industry-wide study that was done and said, “What did we do wrong? How did the polls go wrong?” And I think some of the analysis shows that a lot of the Trump likely voters were not reached as much as we thought by polling, for whatever reason, and there are a whole bunch of explanations. My hypothesis is that it’s maybe related to something more complicated. The reality is that it’s not just Trump. Over the past 20 years or 30 years, Republicans tended to do about two points better on election day than the public polls suggested they were going to be doing.
My hypothesis — and I think I have some evidence for it — is that it was a lot easier for voters to come to the conclusion that they were going to vote for Joe Biden than for the people who were going to vote for Donald Trump. With Joe Biden, it was like, “he’s fine. He’s a centrist Democrat. He’s not the best thing I would want, but the other guy is Trump.” For the Trump voters, [they think,] “the Democrats are going to take my guns, but by the way, I think [Trump] is a horrible human being.” I’ve spoken to people like that, and it’s similar for the issues of abortion, taxes, judicial appointments.
Getting back to the notion of asymmetric politics — there’s something inherent in the way Republicans and Democrats collect their votes that makes it harder for the person who’s going to wind up voting Republican to go “yes, I’m voting Republican.” I think that that’s the essence of it. In the most pejorative terms, we refer to Democrats as pursuing “identity politics” and we refer to Republicans as pursuing something called “wedge issues,” like abortion, gun control, and the like. Inherent in identity politics is the notion that if you’re part of that group, it’s “yeah, that’s right, that’s what I am.” Inherent in the notion of wedge issues is “I don’t agree with them on all this other stuff, but abortion is so important to me; therefore, I will vote that way.” If you’re making those sorts of more complicated decisions, it’s harder for [the wedge issue voter] to be categorized in the right way. And I think that’s part of the reason why Republicans do better on election day than the polling suggests often.
HPR: In an incredibly polarized world, where the differences between sides are very slight, and where it’s difficult to get a high degree of accuracy, what value can politicians and everyday citizens as well, glean from measuring public opinion anyway?
Mishkin: I think you’re asking too much of polling. Polling is inherently inexact. Inexact does not mean inaccurate; it means inexact. I think that you have to assume you are measuring using an old wooden yardstick, rather than a pair of calipers. As long as you look at it that way, the polling can be incredibly informative. This is the sin of all the people who do average — not that averaging is bad — but to the people who just started saying, “let’s look at the averages and build a model around that.” No, you’re assuming stuff when you need to think about the fact that there’s art here, not just science. And I think that’s the important lesson of the past two cycles.
