The unreality of reality TV: How competition shows influence U.S. politics and shape views about economic inequality

Kim: The book is about what happened when America stopped making 60 Minutes the most popular show on television. People don’t believe me when I say it, but 60 Minutes was the No. 1 show on TV as recently as the 1990s. [60 Minutes was No. 1 from 1991 to 1994 and in the top 10 for 9 of the decade’s 10 years.] And then America turned into a nation where American Idol was No. 1 — and that’s the 2000s.

The book is about the political consequences of people getting tuned out of news, because they suddenly had many more choices than they used to. They started to get information about the world not through news, but through entertainment media, particularly reality TV shows, which typically carry with them a dominant narrative about the American dream — the rags-to-riches story of upward mobility.

Kim: And the narrative isn’t as focused on hard work — although I’m sure they all do hard work at the gym to stay looking great.

But I’m talking more about the shows that have, say, a poor waitress from small-town Ohio who has been practicing singing for many years and has stuck with it through thick and thin, and here she is on stage, singing in front of America, living her American dream. There’s sentimental music. There’s a video package that shows the contestant’s struggles — their parents are very sick, maybe, or they lost their home in Hurricane Katrina. Every lemon that life has thrown at you. They talk about those struggles, and the show is the vehicle to turn those struggles into the American dream. This gets repeated across people, across episodes, over and over, and I think it’s really powerful.

Predicting the impact that entertainment TV can have on viewers can be difficult. Like, the same person might watch The West Wing and House of Cards, and those two shows pull our thoughts in very different directions. West Wing is all rainbows and unicorns and politics getting things done and helping people, whereas House of Cards is about corruption in every corner of Washington that could get you killed at any time. Sussing out those differential impacts is very hard. But what was really powerful in the 2000s — particularly the early 2000s — was that so many of the most popular shows on air were reality TV shows that fit this model. It doesn’t matter whether it was Shark Tank or American Idol or The Apprentice or MasterChef — the core narrative is pretty much the same, that anyone who works hard can get ahead and be rewarded.

Kim: Yeah. We didn’t find a lot of effects on voting behavior, because as we know, voting behavior in the U.S. is explained by party ID most of the time. But we did find effects on people’s attitudes, because I think entertainment media can be powerful in affecting how people think about more foundational beliefs in American politics.

So questions like: Are the rich deserving of their wealth? Should we do more to help poor people? Is economic mobility viable in this country. On those we found meaningful impacts. For people who were heavy viewers of these shows, who watched five or six of them, the impact on believing in the American dream was comparable to the effect of having immigrant parents — which is a huge impact. That was really shocking to me, personally.

The other most shocking thing to me was the limited effects of your location — where you live. Raj Chetty at Harvard has done work to quantify the measurable effects of intergenerational mobility. There’s lots of research out there that shows people living in some places have a much easier time moving up economically than people in other places. The New York Times published this huge map that showed all the counties that were high mobility and low mobility. So I was curious whether your location impacted how much you believed the rags-to-riches story.

What I found was much greater effects for the reality TV shows and almost no effect for where you live. The reality on the ground where you live has almost no effect on your actual belief — at least according to the data that I’ve been collecting.

Kim: Yeah — back then, before television, those novels were mainstream entertainment media. And literary critics hated them — they all had the same plot. Why were boys reading these books when there was Shakespeare and Moby Dick? It was too lowbrow, too “popular” at that time. But that’s what people were reading or consuming for entertainment. I think the same goes for contemporary America, where political scientists, or people like you and me and the people we hang out with most of the time, have a tendency to think it’s horrible that people aren’t consuming news all the time to stay on top of the issues of the day; it’s their civic duty to consume news. But literally the first thing you learn in a political science grad program, in your Political Behavior class, is that most people don’t know that much about politics. And people make political decisions, despite having only limited information.

I just found it really intriguing — I didn’t grow up with academic parents, I’m the first in my family to go to college. Who are we to judge someone working five jobs for not keeping up with the news? You don’t have mental or physical energy at the end of the day to know what’s happening at the Supreme Court or the status of some bill in the Senate.

Kim: This is something I always grapple with in my career. When I started studying the political effects of non-political media, it felt really risky. There were a lot of criticisms when I was writing papers about reality TV shows. A lot of people told me “this is not political science research” and that I’d never get a job in academia. But I just couldn’t shake off the feeling that people aren’t interested in politics most of the time, so other things in their environment must be influencing the political decisions they make.

The other thing that helped is the rise of behavioral data. For a long time, we had to rely on surveys to measure people’s individual news habits — you know, a questionnaire that asks “How often do you watch Fox News? How regularly do you read The New York Times?” And everyone answers, “Oh, I read news all the time, I’m a good citizen.” But with behavioral data, we can see exactly which stories people are clicking on and how much time they spend reading or watching. In 2014, the bestselling book in America was Thomas Piketty’s Capital in the Twenty-First Century, this giant 700-page tome of economic theory and data. But if you look at the Kindle data, almost no one actually got past page 20 or so. The behavioral data tells us that most American life is not political — so why don’t we study what people are actually consuming every day, however lowbrow it may seem?

Kim: Yeah, cop shows are another classic example — I’m doing some research on them now. The genre is fascinating because, no matter which one you watch — Chicago P.D., Blue Bloods, NCIS, Law & Order — every episode has basically the same narrative. The criminals get caught in 40 minutes. The cops are extremely effective, and they’re all heroes. If there is some violence, it’s done for the greater good. And maybe there are some bad apples here and there, but the good cops always catch them and filter them out. That’s the narrative about the cops that people are regularly exposed to.

And the people who watch those shows, if you look at the audience demographics, are mostly older, white, and living in a rural or suburban area. Living in New York City, I see NYPD every day. The understanding I have about NYPD isn’t just a function of my media consumption. It’s also the literal cops on the street that I see every day. But when you’re not seeing the police much in your life, the only way you get to know about them is through media. Entertainment media becomes the primary information source.

Think back to 2020, the year of George Floyd, when there was so much news coverage about police brutality. But if you look at the Nielsen ratings for that year, three out of top 10 TV shows were cop shows. Again, I think that’s a prime example where the news media and entertainment media are offering very different narratives, and most people spend a lot more time consuming entertainment media.

Kim: I think it’s mostly money. As the audience gets more fragmented, there are more and more TV shows, meaning there’s less budget for each individual TV show. As a producer, you have an incentive to optimize your budget. And one way to make really cheap TV is a reality show. You don’t need to find expensive actors. There are plenty of people who are very eager to be on your show. And I think Survivor and American Idol led producers to think, “Oh, this is what people want. And it’s cheap. So we should be producing more of it.”

The popularity of these shows has declined a bit since this boom, as people’s media choices have grown even more. Right now, I’m really interested in short-form entertainment — TikTok videos, Instagram Reels, YouTube Shorts. When you’re watching NCIS, you’re immersed in a narrative for 40 minutes. There’s a plot, there are characters you’ve probably come to know over a period of time. Same for many reality TV shows, too. But now we’re talking about entertainment media under 30 seconds — is there a narrative? There’s a micro-narrative. I think we haven’t really thought through what that means for, say, narrative persuasion. I think there’s a lot of room for research in that area.

Kim: Actually, one of my new studies is a huge field experiment throughout the 2024 election season. We recruited around 4,000 young Americans and incentivized them to follow social media creators on TikTok, YouTube, and Instagram for five months. One group was encouraged to follow very political influencers — like, people who talk about politics all the time. The second group was incentivized to follow people who only occasionally talk about politics. They normally talk about cooking, or whatever, but maybe once a month, when they’re cooking, they suddenly say, “By the way, guys, have you guys heard about this housing crisis going on in the Bay Area?”

Our opening question was: Which one is more powerful in persuading people? The ones who talk about politics all day every day, or the ones who mostly talk about other things you care about, cooking or dancing or singing, and then they occasionally bring in politics? Which one’s more effective and more persuasive? The preliminary finding is that the two groups were equally effective — which is really interesting.

Kim: Yeah. But the other is — I think a lot about the role of early education. People don’t read books or news like they used to. We’re losing that reading muscle because of all our TikToks and the dopamine-inducing media content we consume every day. I think we have to do our best to protect the next generation so they have a developed reading brain to process information. So they have some control over what they watch or how they think about it when they grow up.

I spent like yesterday talking to students in my American public opinion class about…what do we do now? Is there anything that we can do as educators, as teachers, or anything that higher education institutions can do to slow down the misinformation or all these challenges we’re having? And a lot of people agreed on the importance of early education — which, of course, also comes with many other challenges around public policy.

But yeah, reading — I think a lot about reading. One of my weird hobbies is that I pay a lot of attention to what people are doing on the subway with their phones. I’m always curious what they’re doing with their phone.

Kim: I don’t, like, look at them up close. But from a normal distance, you can see whether they’re watching something or reading something or playing something. In my entire last year of observation, I’ve only seen two people reading news. Two, out of all the people I’ve seen on all the subway rides I’ve taken.

It’s a small number, but it’s consistent with the behavioral data that we see. It’s a very small subset. Most people are texting, watching YouTube videos, scrolling Instagram, looking through their photos, playing a game — occasionally, reading a book. But that’s also rare. I always tell my students: Go out and look at what people actually do. Think about the questions you can answer based on real-world observations, not ivory tower expectations or what you read from a pundit in the Times. Talk to real people.

I always ask my Uber driver to guess how many people in America regularly watch Fox News. You can see I’m a very annoying Uber passenger.

Kim: 50% of America. They think half of Americans are Republicans, and they all watch Fox News. [In Q1 2025, Fox News averaged 3 million viewers a night in primetime, less than 1% of America’s population. In “the demo” — people aged 25-54 — it averaged only 361,000 a night.]

The partisan media effects are very interesting. If only 3 million people are regularly watching Fox News — but if most people think that half the country is — that perception has huge consequences on polarization and how we think about the other. I think that’s where the news media effects get amplified — not because a lot of people are watching, but because it changes how we think about what others are up to, or what others are doing.

Benton: Let me circle back a bit. You said that the biggest impact you found in your research was in attitude change, as opposed to voting behavior change. But I also know you had that paper about The Apprentice that found people watching the show had a meaningful impact on Trump’s success in the 2016 Republican primary. What was different there?

[From the paper’s abstract: “Using an array of data — content analysis, surveys, Twitter data, open-ended answers — we investigate how this TV program helped Trump brand himself as a competent leader and foster viewers’ trust in him. Exploiting the geographic variation in NBC channel inertia, we find that exposure to The Apprentice increased Donald Trump’s electoral performance in the 2016 Republican primary.”]

Kim: That paper was tricky, because there was not one ideal data source that I could rely on. It was a mix of a bunch of different data sources collected in one place to figure out whether there was a compelling story there — polls, transcripts data, Twitter data. I was looking at people back in 2004, when The Apprentice first aired and was very popular. At the time, we didn’t have that many TV choices. A lot of people used to stay on one channel — meaning that if you watched Seinfeld, you watched whatever NBC show came on after Seinfeld. I was looking at the ratings for the very popular TV shows that aired right before The Apprentice — Joey and Will & Grace, depending on the season. I was looking at the proportion of people who watched the 8 p.m. show and then stayed to watch the 9 p.m. show. I was basically quantifying the effects of people who were not selecting to watch The Apprentice, but who happened to watch The Apprentice because they were watching the show before it.

One thing that I found really, really convincing about the analysis, that made me more confident about the result, is that I didn’t find any similar effects for any other Republican primary candidates in any other elections. And I didn’t find the same effects in the general election, where party ID was the big factor, along with whether you hate Trump or Hillary. The fact that I didn’t see effects in the general made me more confident that the effects in the primary were real. You’ve got 10-plus candidates up on a debate stage, you don’t know who these people are, and name recognition matters a lot in a primary.

Kim: So the interesting context for that paper is that, in political science and other fields, we have tons of papers built around doing experiments — meaning we randomly assign some people to a condition, other people to nothing, and we try to measure if the condition had an impact on whatever we’re studying. Like, we randomly tell people to deactivate Facebook for a month, or watch Fox News for a month, or whatever.

And one of the biggest problems in that approach — and I say this as a person who runs experiments, too — is that, in the real world, no one watches TV because researchers tell them to. That’s not real life. Many of the papers we see in our field on, say, the effects of watching local news or Fox News or whatever are based on these experimental methods. And that experimental method of forced exposure is the direct opposite of the actual media environment right now, which is defined by choice. We choose what to watch, what to read. But the primary method we use in social science is to not give people any choice.

The paper you’re talking about is an attempt to try out a different methodology that we think might reflect the real world better — to both force them to watch and to give people the choice to watch, and then compare the effects.

[The paper in question asked people in five African nations — Botswana, Cameroon, Nigeria, Uganda, and Zambia — to watch some mix of Chinese-produced videos about the wonders of China’s economic system, American-produced videos about the wonders of the U.S. system, and nature videos, which served as the control. Some had no say in which videos they got to watch; some were allowed to say whether they wanted to see the China, U.S., or nature videos. The results? “Propaganda’s effects are especially large for those who would ordinarily not choose to watch it” — meaning the people who didn’t want to watch pro-China or pro-U.S. videos were more influenced by those videos than those who did.]

Our finding was the effects were much more pronounced on the people who were not likely to choose the content if we give them a choice. That raises a whole lot of questions, and I think it merits a little bit more skepticism in how should we think about many of those papers using experiments. In real life, if no one’s watching it or being exposed, whatever we’re estimating is just a counterfactual.