Why Is U.S. Media So Negative? (Ep. 477)

Breaking news! Sources say American journalism exploits our negativity bias to maximize profits, and social media algorithms add fuel to the fire. Stephen Dubner investigates.

Listen and follow our podcast on Apple Podcasts, Spotify, Stitcher, or wherever you get your podcasts. Below is a transcript of the episode, edited for readability. For more information on the people and ideas in the episode, see the links at the bottom of this post.

*      *      *

I’d like you to imagine — and this shouldn’t be very hard — but imagine you’re in the midst of a growing pandemic:

A.B.C.: A twelfth fatality has been reported here in the United States.

N.B.C.: E.R. doctors saying we are on the verge of a medical disaster.  

And let’s say you want to be as informed as possible.

C.N.N.: The daily coronavirus death toll in the United States might hit 3,000 by early June. 

N.B.C.: The coronavirus forcing millions more Americans into virtual lockdown. 

Fox News: This dangerous health crisis could dovetail quickly into a political crisis.

And now let’s say you are an economics professor watching this news, for hours a day. How does the information you’re getting add to your understanding of the pandemic? 

Bruce SACERDOTE: I honestly thought I was going crazy.  

The economist in question here is Bruce Sacerdote, at Dartmouth College.

SACERDOTE: I’m very utilitarian, and I was looking for useful information and hence my frustration because I felt like it was more advertorial and entertainment.  

When Sacerdote says he was looking for “more useful information,” what does that mean?

SACERDOTE: What I would be looking for is, “Okay, there was this new study done, here’s what they found. Here’s what this means for the pandemic. Here’s what this means for when we can get back to work.” But instead, it tends to be a lot of angst and bemoaning the numbers, even if they hadn’t changed or had gotten better.

It wasn’t that Sacerdote wanted to pretend that everything was fine.

SACERDOTE: I mean, this thing killed more people than most of the wars we’ve been in. It’s hard not to be knocked down by that. 

But Sacerdote saw a difference between being knocked down and wallowing. He began to wonder if the news coverage of the pandemic was commensurate with the pandemic itself. And whether the coverage he was seeing — mostly from major U.S. media outlets — whether it was perhaps more negative than other coverage. Like local news, or international news, or even the articles published in scientific journals. All of them were seeing the same Covid-19 story unfold — but were the major U.S. media outlets selling a more negative version of the story? And if so, what were the ramifications? Sacerdote wasn’t quite alone in his concern. The Centers for Disease Control issued a warning about media consumption: “Take breaks from watching, reading, or listening to news stories, including those on social media,” the C.D.C. said. “It’s good to be informed but … consider limiting news to just a couple times a day.” Now Sacerdote, remember, is an economist — not an epidemiologist or a public-health scholar. So he also wondered how the economic setup of the U.S. media industry was driving the tone of the coverage. We have been putting out a series of episodes lately about how the U.S. is fundamentally different from other countries. We’ve looked at the extraordinarily high levels of individualism in America; we’ve looked at policing and child poverty, transportation policy. Did Bruce Sacerdote find another dimension on which the U.S. is an outlier? He decided to do what economists do; he started gathering data in order to produce a study. 

SACERDOTE: That’s what happens when you take an economist and lock them in front of C.N.N. for three months and make him more and more angry. 

Today, on Freakonomics Radio: is your news negatively biased? Or should we just blame the English language?

Arika OKRENT: We have a lot of words for types of bad feelings.

And let’s not forget about social media.

Steve RATHJE: We found that moral words, like “evil” or “hate,” these would also be linked to increased virality. 

Marshall McLuhan said it first, more than 50 years ago: “The medium is the message.” It’s also been said that we call T.V. a medium because it’s neither rare nor well-done. Is that fair? Is it true? Is it time to answer these questions? Yes, it is.

*      *      *

In certain circles — like academia, where Bruce Sacerdote works; or journalism, where I do — you are generally considered a more serious person if you are critical, or even negative. Whereas positivity tends to be associated with naiveté or cheerleading.

SACERDOTE: Yeah, that’s probably right.  

So Sacerdote, who is generally an optimist, sometimes feels like an outlier.

SACERDOTE: I had this idea of getting together a merry band of people who actually believe that there’s economic growth, and that poor people are becoming better off in the U.S., even if not as fast a rate as rich people. And I wonder if having people like that band together could be more effective than one or two voices crying in the wilderness. It’s interesting that when it comes to the business world and particularly the high-tech world, I think there is plenty of optimism. And certainly investors — I mean, you look at companies like Tesla and Amazon that exist in part because investors subsidize their operations with the optimism that 10 years down the road, the big payoff is coming. So there’s an interesting dichotomy there, right? 

It is true that tech investors can be incredibly optimistic, sometimes to a fault. How optimistic should you be if you’re investing in media firms? It depends. Newspapers are, for the most part, a bad bet. Over the past 15 years or so, the digital revolution has upset what used to be a very profitable apple cart: U.S. newspaper revenues fell from around $60 billion a year to $20 billion. But cable T.V. is doing great. At the big three — Fox News, C.N.N., and M.S.N.B.C. — revenues continue to grow, and their profit margins are massive. C.N.N., for instance, earned an estimated $700 million profit in 2019 on revenue of just $1.6 billion.

SACERDOTE: It’s not only a for-profit enterprise, but it’s highly profitable and it’s a big market they can segment.

And while technology markets thrive on optimism, Sacerdote began to suspect that major media outlets thrive on pessimism.

SACERDOTE: The media is very good at producing negative stories that are eye-catching. They’re also good at producing stories which cater to people’s existing tastes and fears.  

Fear-mongering, we should say, is not new. Journalism does span a wide spectrum, but the most crowd-pleasing outlets have long followed a simple mantra: if it bleeds, it leads. And Sacerdote argues that this instinct is particularly strong in the U.S.

SACERDOTE: Essentially, the U.S. major media is better at giving people what they want. They’re particularly talented and profit-maximizing.

How did he reach this conclusion? It was the result of a huge research project done in collaboration with Molly Cook and Ranjan Segal. They set out to analyze Covid news coverage in four distinct categories: major U.S. media, local and regional U.S. media, international media outlets, and scientific journals. All told, they analyzed 43,000 stories, including journal and newspaper articles and cable T.V. transcripts. They used machine-learning algorithms and what Sacerdote calls “very simple word-counting techniques” to measure how negative or positive a given story was. This measurement relied on the use of two lexicons popular with researchers — a list of nearly 5,000 words judged to be negative and another list of positive words, just over 2,000 of them. Perhaps the difference in size between these two lexicons should have been a clue. Here are some of the words from the negative lexicon:

ANNOUNCER: Appalling, barbaric, catastrophe, dangerous, nefarious, recklessness, stagnate, troublesome, worsen.

And, some words from the positive lexicon:

ANNOUNCER: Applaud, appreciate, expansive, gaining, ready, winnable. 

The researchers focused their analysis on Covid coverage because that’s what Sacerdote was interested in. But the particulars of the pandemic also allowed them to sharpen their analysis, since the virus hit different places at different times. This meant they could use local Covid trends as a control tool to isolate and measure the tone of the media coverage. So, what’d they find?

SACERDOTE: It doesn’t seem to be driven by the trend in cases that much. And that’s really disturbing. Because when the news is terrible, I expect terrible reporting. But we counted up the number of negative and positive stories, both in times when cases are rising, and when cases are falling. And when cases are rising, negative stories outnumber positive ones seven-to-one or six-and-a-half-to-one. And then when cases are plummeting, it’s still five-and-a-half-to-one negative stories to positive stories. I mean, that’s really upsetting. 

DUBNER: So, you write that about 87 percent of Covid coverage in national U.S. media last year was negative. The share for international media was only 51 percent. So that’s a massive, massive difference. Only 53 percent in U.S. regional media. So, again, huge difference. And then 64 percent in scientific journals. So that’s really interesting to me because, I guess it has to do with the difference in mission between journalism and scientific journals. What do you make of that headline number: 87 percent negativity in national U.S. media versus 64 percent negativity in scientific journals overall?

SACERDOTE: Of course, it’s not proof positive because, as you say, these entities have different missions. But to us, it’s pretty astounding. It’s like, look, yeah, there’s bad news about Covid. And maybe some of those scientific articles are about spread and those sorts of things. But on the other hand, there’s a good 40 percent that are finding good things. The positive scientific news just doesn’t get out there. And often when it gets out there in the mainstream media, it gets kind of botched. 

For example: think about how the vaccine timeline was covered early on. Here is a New York Times headline from April 29th, 2020:

ANNOUNCER: “Trump Seeks Push to Speed Vaccine, Despite Safety Concerns.”

And here’s a passage from that article:

ANNOUNCER: “President Trump is pressing his health officials to pursue a crash development program for a coronavirus vaccine that could be widely distributed by the beginning of next year, despite widespread skepticism that such an effort could succeed and considerable concern about the implications for safety.”

Now, we should say, the New York Times’s coverage of President Trump was almost uniformly critical, so maybe it’s not surprising that its coverage of the Trump administration’s vaccine efforts might also be critical. But the Times was just one of the 14 major U.S. media outlets in this analysis. Here, meanwhile, is a headline from one of the foreign outlets the researchers analyzed — this is the Oxford Mail in England, in February of 2020, also writing about vaccines:

ANNOUNCER: “Scientists Working on a Coronavirus Vaccine in Oxford.”

And, a passage from that article:

ANNOUNCER: “The Jenner Institute has been working on a vaccine against another coronavirus, Middle East Respiratory Syndrome (MERS), which has been shown to induce strong immune responses against MERS after a single dose of the vaccine in the first clinical trial, which took place in Oxford. The same approach to making the vaccine is being taken for the novel coronavirus vaccine. 

So this is the kind of information that Bruce Sacerdote the economist would see, later, in his analysis. But at the time, Bruce Sacerdote the human was back in Hanover, New Hampshire, watching a lot of C.N.N. and reading a lot of the New York Times. 

SACERDOTE: I literally thought I was losing my mind, and I’m thinking, well, why am I putting faith in these scientists who say they can come up with a vaccine because the news clearly says they can’t. That it takes five years to develop a vaccine. And then when the scientists and the companies actually came out with vaccines, I felt so relieved and, less importantly, vindicated that, hey, wait a minute, the scientists were not lying. It’s just that they weren’t being given a full hearing. 

DUBNER: In the paper, you write that the most popular stories in The New York Times have high levels of negativity, particularly for Covid-19 related articles. So is The New York Times, just to use one example, more negative than others? 

SACERDOTE: The data suggests that The New York Times is more negative than the average regional or local paper or T.V. 

DUBNER: And what can you tell us about which way the arrow points there? In other words, do New York Times readers want and seek out negative news? Or does The New York Times turn people negative? 

SACERDOTE: That’s a great question, and I’m sorry to say we don’t have the answer. It would almost be less worrying if people demand negativity, and The New York Times supplies it. And we’re picking on The New York Times, but that name stands in for all the U.S. major media. It would be less disturbing if it were simply that people demand negativity, they get it, the end. What worries me is that it’s a self-reinforcing cycle.

In other words: it could be that the bad news delivered by major U.S. media outlets increases our appetite for bad news, and in order to maintain its audience, those outlets in turn deliver even more bad news. To pick on The New York Times just a bit more — in a separate analysis, the data scientist Kalev Leetaru performed a sentiment analysis on every article the Times published between 1945 and 2005. He found that coverage began drifting negative in the 1960s and has gotten progressively more so. But, which way does the arrow point? Do news outlets simply meet our demand for negativity? Or do they create that demand? It’s been well-documented by academic researchers that humans do have a built-in negativity bias. The social psychologist Roy Baumeister calls it “the power of bad” — and he says it can serve a valuable function.

Roy BAUMEISTER: If you miss out on a great opportunity for good food or sex or any other life-affirming thing — well, okay, that’s too bad. But you might have another one the following day. But if you miss out on a dangerous predator, fail to notice, that will put an end to your life. Part of the psychological mechanism underlying our work is that the mind was shaped by evolution to pay attention to risk.

So how does the “power of bad,” an ancient psychological mechanism, intersect with how The New York Times conveys information? To understand that, it helps to first understand how the English language has been shaped by this negativity bias.

OKRENT: It’s certainly the case in language that negativity heightens emotion, and it heightens the impact of what you’re saying. 

That is Arika Okrent.

OKRENT: I’m a linguist, I write about language, and I’m the author of Highly Irregular, about why English is so weird. 

“Weird” in some simple, relatively harmless ways — like spelling. One example from Okrent: consider the following three words: “D-O-U-G-H,” “T-O-U-G-H,” and “T-H-R-O-U-G-H.” Other than the opening letters, they’re identical. So why don’t they sound identical? Why are they pronounced “dough,” “tough,” and “through”? But the weird spelling in English isn’t nearly as complicated as the weird emotions. Especially the negative ones.

OKRENT: We have more ways of being negative. We have a lot of words for types of bad feelings. We have “guilt.” We have “shame.” Those are very specific. They’re different from “sad.” They’re different from “down” and “depressed.” And there are definitely lots of ways of being positive or happy. But the vocabulary we have for it, there aren’t very specific ones that are like “that particular type of joy you feel when you sit down to a meal that looks really good,” or something very specific like that. 

DUBNER: I have to say, this is a little distressing to me because one of the most famous lines in literature is the first sentence of Anna Karenina by Tolstoy: “Happy families are all alike; every unhappy family is unhappy in its own way.” I’ve always thought that was such a bad caricature, that happiness must have as much variety as unhappiness. But you’re telling me, basically, no surprise — Tolstoy’s right, I’m wrong.

OKRENT: Well, you can definitely describe different types of happiness, but we tend to do a long description to get exactly what we’re talking about. We don’t have the word that sums it all up in one package. There’s a default, and this is how we expect things to be, and then there’s a marked situation. And in language, the marked situation is the one that gets the name because it’s a departure. Especially with verbs. We have the verb “to lie,” “lying.” We do not have a verb for “to tell the truth.” We have a whole phrase. We can say, “Oh, he’s being a straight shooter.” But we have one word for lying. You know, what’s the opposite of speeding, of littering, of murdering? That’s what we’re supposed to be doing. We don’t need to name it. We’re just going along with it. 

DUBNER: When we consume media, how do you think the negativity of the headline language shapes our perception of the events?

OKRENT: It shapes how we react to the events and how urgently we feel that reaction. Negativity puts you in a heightened state of awareness, and that heightened state of awareness is meant to spur you to action.  

The music also helps. You know, the music that C.N.N. and other cable-news networks play to make sure you know that their “Breaking News” alert is really important.

ANNOUNCER: Breaking news: Christians celebrate Christmas.  

These are real examples from C.N.N.; the “breaking news” text is being read aloud by our producers.

ANNOUNCER: Breaking news: No winner yet in America’s historic election.

ANNOUNCER: Breaking news: Titanic Sunk 102 Years Ago Tonight.

We should also acknowledge that a lot of news is meant to alarm us; that’s part of its purpose. And I know journalism is a business. It’s the business I’ve been in most of my adult life — including several years at The New York Times. But I’ve always thought of journalism as having a somewhat different mission from other industries. Yes, every writer and editor and producer wants their work to get attention, and they want to be paid. But the argument being made by Bruce Sacerdote goes beyond that. He says that the major American media outlets are primarily driven by profit-maximizing, and that the best way to profit-maximize is by accentuating the negative.

SACERDOTE: Yeah, that’s the sad truth. I do think that the realm you’re in, the profession you’re in, people expect a certain level of truth. And I think that they often get that. And so I feel that some of that trust that had been built up over, let’s say, 100 years is partially eroded because it is perhaps more of a business than it was in the 1950’s. 

DUBNER: But in retrospect, doesn’t that look like a brief golden era? Because if you go further back to the late 19th century, the papers were —.

SACERDOTE: Oh, yeah, they were just rags. They were political rags, and they were very explicit about it. If they were writing a paper about the British or about George Washington, it was just a personal attack. So, for better or for worse, it’s a business and people are getting what they want. 

DUBNER: Do you know anything about how American the taste for negativity is, and assuming it is an anomaly, why that’s the case? 

SACERDOTE: We take the opposite view on this, which is that it’s not actually the people that are different. And I’ll give you some data on that. If you look at the most Facebook-shared and the most-liked stories on, say, The New York Times or the B.B.C., the most-liked things from the B.B.C. are also super-neg. It’s just that the B.B.C. is not supplying nearly as much of those super-negative stories. What we, my coauthors and I, think is going on is it’s not that Americans are fundamentally different than the British or the French or the Italians. We think what’s going on is the structure of the industry is different in these different places. The U.S. major media outlets explicitly focus on the negative because, we believe, that’s what drives viewership and clicks and keeps people staying on the page or on the show. 

DUBNER: Okay, but why would that not be the case in France or England or Australia? 

SACERDOTE: In most of those other countries, you have a big public player like the B.B.C. or the Canadian Broadcasting Corp. As a libertarian-leaning person, you think I wouldn’t be pounding the table for public interference in this industry. But I think in this industry, they have less of a profit motive. They are somewhat less motivated by driving clicks and engagement and somewhat more motivated by the truth. 

Sacerdote’s study covered just 2020, from January 1st until December 31st. So, it did include the beginning of the vaccine rollout, but it ended before Joe Biden became President. Sacerdote did look for a relationship between the political bias of a given news outlet and its tendency to run negative news; he didn’t find any. But you could imagine that Donald Trump’s contentious presidency may have affected the overall tone of media coverage in 2020.

SACERDOTE: I certainly feel like the negativity is somewhat less pounding than it was six months ago, even in the face of this horrible rise of this variant. So, then it does make you take pause and say, well, maybe part of it was the political environment we were in.

But Sacerdote thinks his research findings are more generalizable than that.

SACERDOTE: We suspect this is much more than just a Covid story. We think that negativity about climate change — pick any topic: unemployment, inequality, poverty alleviation — we think that the media coverage, particularly from the national media, is probably more negative than it is in other countries. 

Why is this important — other than the psychic damage that so much negativity can cause? Here’s one reason: if all you’re being told by the media is that Problem X is bad and getting worse, and Problem Y is even more unsolvable — well, you may start believing it. You may start believing that we are collectively terrible at solving problems, and it’s probably not even worth trying. Whereas the reality is that collectively, we humans I mean, are actually quite good at solving problems. Yes, it’s hard — but it’s made even harder when the only stories that gain traction are the stories telling us that those problems can’t be solved. Now, I know what you’re thinking. You’re thinking — wait a minute, there is one newish sector of the media that’s practically devoid of negativity. Social media. Everything on social media is puppies and rainbows, isn’t it?

RATHJE: One example of a very viral post said, “Check out Joe Biden’s recent brain freeze.” 

*      *      *

There’s a famous saying in poker: When you’re sitting at the table and you can’t tell who’s the sucker, the sucker is you. Here’s another version, updated for our digital age: If you’re spending a lot of time online and you can’t tell what the product is, the product is you.

RATHJE: The entire social-media business model is based on capturing our attention in order to sell advertising.

That’s Steve Rathje. He’s getting his Ph.D. in psychology at the University of Cambridge. He studies misinformation and political polarization. He also created a Web app called “Have I Shared Fake News?”

RATHJE: You input your Twitter handle and we’ll show you specific news U.R.L.s that you have shared that have been considered by independent fact-checkers to share fake or unreliable information. 

As Rathje said, the big social-media sites are almost exclusively reliant on advertising dollars. In 2019, Twitter took in just under $3.5 billion in revenues; that same year, Facebook took in $70 billion. This means that all the newspapers in America, even all the cable-T.V. networks, could fit in Facebook’s back pocket. And in a way, they do: More than half of all Americans get at least a portion of their news via social media, with one-third coming from Facebook. Here’s what Steve Rathje wanted to know: if you’re a social-media site and your business is built around engagement in order to sell the most advertising possible, what’s the best way to drive engagement? So Rathje, like Bruce Sacerdote, embarked on a big study. He and two co-authors — Sander van der Linden and Jay Van Bavel — analyzed nearly three million social-media posts to learn what makes a post more likely to attract other users. Their analysis covered the years 2016 to 2020; they focused on posts from conservative and liberal media platforms and Republican and Democratic members of Congress. So, what’d they find?

RATHJE: What we found is that each additional word referring to the outgroup increased the number of retweets or shares of that post by 67 percent. 

The “outgroup” meaning someone on the other side of the political aisle.

RATHJE: If a post was coming from a Democrat, a word like “Republican” or “conservative” would lead to increased virality. And if a post was coming from a Republican, a word like “Joe Biden” would lead to increased virality. 

The paper, which was published in the Proceedings of the National Academy of Sciences, is called “Outgroup Animosity Drives Engagement on Social Media.”

RATHJE: Also, posts, including words referring to an outgroup, were also much more likely to receive angry reactions, ha-ha reactions, comments, and shares, whereas posts referring to an ingroup were more likely to receive like or heart reactions on Facebook. But posts about the outgroup receive much more engagement in total. 

DUBNER: Steve, how confident are you that you and your coauthors are right? In other words, how empirical is this kind of research? 

RATHJE: We looked at eight separate data sets on Facebook and Twitter, so we are confident that we are right, especially about this specific point in history. Our general results were also replicated. 

DUBNER: Can you describe the two most viral posts in your entire, massive dataset? 

RATHJE: One example of a very viral post was from Breitbart News, and it said, “Check out Joe Biden’s recent brain freeze.” And it was a very unflattering video of Joe Biden that made him look like he wasn’t doing so well. And then another post that went very viral from the liberals was from The Daily Beast, and it was about Mike Pence blatantly lying about Covid-19. 

DUBNER: Let me just make sure I understand, Steve. What you’re telling me is that when I tweet or post something on Facebook or pretty much anywhere, if I want to be successful — because here I am posting, I’m not here to be invisible — all I really need to do is focus on my outgroup and being negative about them, and I win, correct? 

RATHJE: Yeah, you might win in terms of engagement. You might not get people to like you. But if you have a rival podcast, if you wanted to dunk on the rival podcast, that would probably get you a lot of engagement.

DUBNER: So, in my next series of tweets, I should say that Ira Glass’s glasses are too big, and he looks like a circus clown?

RATHJE: You should try that. We should put that to the test. 

DUBNER: Can you help me get better at it? Like, what are some kinds of words or emotions or actions to post about? 

RATHJE: Well, let’s see. If we take from other research — so I was inspired by other research that also looked at the effect that moral outrage would have on virality. And we also replicated this effect as well. We found that moral words, like “evil” or “hate,” or they could even be positive moral words like “care,” these would also be linked to increased virality. So maybe if you expressed some moral outrage about Ira Glass.

DUBNER: Yeah, I feel like Ira, like he is pretty nice. What about maybe — so Joe Rogan is a natural target as a competitor, but he could also just beat the sh*t out of me with with one finger. So that’s not a good idea, is it?

RATHJE: Well, if you guys got into a feud on social media, that would probably drive a lot of engagement as well. 

DUBNER: I’m okay with the feud on social media, but if it tips over into real life, I’m dead.

RATHJE: That’s true. When we published the study, we really wanted to make sure that this didn’t come across as advice for people. We wanted to emphasize that this reflected the perverse incentives of social media.

These “perverse incentives,” as Rathje categorizes them, are not universal. Just as journalism operates under different guidelines around the world, so too do Twitter and Facebook. In China, for instance, social-media content is tightly regulated, especially any posts about politics. Twitter is outright banned in China, although many people use virtual private networks to get around the ban. Some Chinese Twitter users have been jailed for criticizing the government; and during the Covid-19 outbreak in Wuhan, the government clamped down on social-media activity that documented what was happening in the city’s hospitals. In the U.S., meanwhile, the government has been pretty much absent in regulating social-media activity. The occasional high-profile banishment of a user like Donald Trump has come from the companies, not the government. But that may change, as politicians on both sides have been calling for more regulation.

RATHJE: Conservatives think that they’re being censored. Liberals are more concerned about misinformation. But we found that outgroup negativity was equally likely to go viral for both Republicans and Democrats, and it was also equally likely to go viral on Facebook and Twitter. So I think one potential solution that both the Republicans and Democrats could agree on is maybe we just shouldn’t amplify this extremely negative content about our outgroups all the time. 

DUBNER: Imagine we’re going back a couple thousand years and you’re saying to a bunch of Roman senators — because they were at least as contentious as modern politicians, probably more so. And it’d be like going back to them and saying, “Listen, you guys just shouldn’t say negative and especially mean things about the other people in this arena.” Does that seem remotely realistic in whatever millennium we’re talking? 

RATHJE: I think that what social media is taking advantage of is an ancient instinct to pay attention to the negative, or to the polarizing or divisive. But I think it is different now that there is the ability to algorithmically amplify this. They didn’t have these outrage machines at that time that would just amplify the most negative content.  

DUBNER: What is the payoff of all this attention? What do the politicians, for example, actually gain? I guess I’m asking you to prove that this virality has real, measurable value.

RATHJE: What we see is a lot of the most extreme politicians — if you look at Donald Trump, for instance — he was really good at taking advantage of Twitter to get the spotlight constantly.

DUBNER: I mean, that’s the narrative. But he also lost an election as an incumbent president, which isn’t easy to do. So, of course, one narrative could say: well, he was great at social media, and he was president because of social media. Another narrative could say: because he was so hostile and negative, he lost an election. So that’s what I’m asking, because attention for attention’s sake isn’t necessarily the goal. So, is there a way to actually measure the value of dunking on someone else? 

RATHJE: That’s a good question, because certainly Joe Biden, he’s not a big Twitter user. You could certainly take that perspective as well because dunking is certainly a double-edged sword. You will get yourself more attention, but also get yourself negative attention. There was a paper that shows that people don’t really like when politicians are negative. Politicians can get more visibility, but they’ll also be perceived as more unlikable, so it is sort of a game that they have to play. 

DUBNER: And what do you see or know about this same phenomenon, outgroup negativity, in totally nonpolitical realms and even non-media realms. Let’s say it’s one athlete on one team dunking on somebody on a different team? Or what about commercial products? If this phenomenon is so powerful, why is Coke not just constantly trashing Pepsi? When I look at the Coca-Cola official Twitter account — let’s see, how many followers does Coca-Cola have? 3.3 million. So they’re doing okay. Although I would think, Coca-Cola, you could do more. And if I load the tweets back to November of last year, 2020, and I search their timeline for the word “Pepsi,” I get zero. They’re not engaging at all. So, if I were Coca-Cola and I’m listening to Steve Rathje, I would say, “Holy crap, we’ve been wasting this amazingly great opportunity to tell the people who love us that, I don’t know, Pepsi-Cola puts rat tails in their soda.”

RATHJE: Yeah, again, I’m not giving people the advice that everyone should go dunk on their outgroup right now. I do think that if they did do that, they would get more attention on Facebook and Twitter, but they might not get people to like them more. Coke has all these very positive commercials. They associate themselves with a positive image. So, yeah, it’s not the best tactic for getting people to like you. It is just a rule of social media that will get you more attention. 

SACERDOTE: I can’t blame the platforms directly.  

That, again, is the economist Bruce Sacerdote.

SACERDOTE: Certainly, one can blame social media platforms for allowing completely false things to circulate. And so you can have a debate about that and the degree to which that might be regulated. But if there’s a negative story in The New York Times, then people are going to want to share it on Facebook. And I mean, that’s what the data show: people do.

According to Steve Rathje, that is pretty much the exact argument that companies like Facebook and Twitter make when they’re accused of using their algorithms to promote negative or even false information.

RATHJE: They essentially argue that social media is a mirror of society. 

But Rathje isn’t persuaded. 

RATHJE: Our research would suggest that social media amplifies the bad and it amplifies the ugly. And the good has a lot less of a chance of going viral. Social media isn’t just this neutral public square in which people have debates. The most divisive or negative content will capture our attention. 

DUBNER: I guess if I’m Facebook or Twitter, though, I could say, “Well, both can be true in that we are a platform where people can say pretty much whatever they want to say — good, bad, neutral. And then it comes down to preferences and what people actually want.” So what’s your evidence that the platforms are actually guilty of accentuating the negativity? 

RATHJE: You could say that we want negativity because negativity is more likely to capture our attention, but I don’t think that people actually want negativity in the long run. Some evidence for this comes from other research. There was another study by Hunt Allcott, in which he paid people to delete their Facebook accounts for four weeks. And after those four weeks, people became less politically polarized and they actually reported better well-being. When we leave these platforms, we often are happier.

DUBNER: I know you said you didn’t write this paper to give advice for how to go viral on social media, but do you have suggestions for how to change the incentives that create this phenomenon? 

RATHJE: So I think that Facebook could make very subtle algorithmic tweaks to just make it so angry reactions cause less virality — and perhaps heart reactions and “like” reactions lead to more virality. There is also other research, by Katy Milkman for instance, that shows that high-arousal, positive emotions are likely to go viral as well. So if the algorithm was subtly shifted, so we take advantage of viral positivity rather than viral negativity, that might be a potential solution.

DUBNER: That makes so much sense to me. We saw not so long ago the Summer Olympics, from Japan. When you watch the coverage of the Olympics, it’s almost as if negativity is barred. Especially if there’s an American favorite who ends up doing very poorly, you pretty much never hear about it. When you see the features on the athletes, there’s always the negative — but it is just the barrier which the athlete overcame to get to triumph. So if that’s the tenor of coverage of an event like the Olympics, which is a pretty big global event and which is hugely profitable, why on earth wouldn’t I think that positivity has a lot of value and that negativity maybe is exciting and fun, but kind of a loser’s trap. 

RATHJE: Yeah, I don’t know. 

DUBNER: Sorry, that wasn’t really a question as much as a sermon. I apologize, but if you have anything to say about it —. 

RATHJE: We should make social media more like the Olympics, yeah! And I mean, there are some platforms — I think TikTok early on was taking advantage of viral positivity. And I think like YouTube in its early days, was a lot of very positive, uplifting videos. But recently we’ve seen like more controversies about the YouTube algorithm recommending conspiratorial or white-supremacist videos. So I think it’s a product of social media evolving with this business model of just constant engagement all the time. And, yeah, the power of bad is a strong bias. 

The power of bad may indeed be a strong bias, but remember, it’s not equally powerful in all domains. Bruce Sacerdote again:

SACERDOTE: The data suggests that The New York Times is more negative than the average regional or local paper.

Here, again, are the negativity numbers from the Sacerdote media study: 87 percent of Covid coverage in national U.S. media, like the New York Times, was negative. The negativity number for regional and local coverage: just 53 percent. So, maybe there’s room for some optimism? Probably not.

SACERDOTE: Local newspapers do tend to go broke at an alarming rate. 

Since 2004, one in five local newspapers has shut down.

SACERDOTE: So they may just simply not be as profit-maximizing. Another possibility is that they play a completely different role, that they focus on local happenings. And so when there’s a fire, which is a very negative thing, you tend to get a lot of reporting about it, but they’re not in the same business of getting people whipped up into a frenzy and getting a lot of clicks and attention — and are less successful as a result.

So this sounds like a losing formula: If you are a media outlet that doesn’t promote negativity, you’re more likely to go out of business.

SACERDOTE: Yeah, it’s a huge issue. There ought to be a market force for local and regional coverage. I mean, the good news is that we have all these new technologies for reaching people. We have all these less-expensive ways to get the word out there. And so maybe 10 years from now, we’ll come back and the media industry will have realigned itself and maybe for the better.  

Sacerdote has already admitted to being an optimist — so this may just be the optimism talking — but he does see an upside in media coverage that doesn’t just bang on about a problem, but instead looks at the problem from multiple angles, and maybe even explores a solution. For example?

SACERDOTE: Look, there’s all this vaccine hesitancy. Perhaps some of the vaccine hesitancy is actually because there wasn’t as consistent a positive message about the vaccines, about the folks that were developing these vaccines, and the public-private partnerships that created them with such speed. I certainly can’t say that a huge fraction of vaccine hesitancy is down to that, but maybe some of it is. I think the entire attitude of the country towards the fight against Covid, but even the fight against inequality, against poverty, and the whole view of whether government works — I think we’re way too pessimistic. I think we’re way too pessimistic about our ability to fight climate change and to get off of fossil fuels. 

C.N.N.: The alarm bells are deafening.

SACERDOTE: We’re way too pessimistic about our ability to beat back Covid and the next virus when that comes along. 

M.S.N.B.C.: Americans are literally too dangerous to be let out of our country.

SACERDOTE: And we’re way too pessimistic about our ability to get people out of poverty. 

Fox News: This entire thing is and always has been a scam.

SACERDOTE: It just bugs me. And I think this negativity is holding society back, rather than looking at what we can do as opposed to what we’ve done poorly. There should be a market for sensible moderates who believe, “Yeah, government can work, but it’s not always the solution. And maybe we should care about the deficit, too. Covid is bad, but look at all the great things we’ve done.” But oddly, you don’t see the Moderate Broadcasting Corporation starting up and gathering viewers like crazy. Although maybe, in a way, that’s what you do. I know that’s not exactly what you do, but you’re kind of in that space. 

DUBNER: I was going to say I’m a little insulted, because you’re describing exactly what we try to do. On the other hand, you’re right: I do not have the scale of what would be called the Moderate Broadcasting Company. So maybe a name change is in order. But basically you’ve told me, if I have any self-interest at all, I’m being an idiot by not being much more negative. Do you have any ideas for names if I want to go the other way? Maybe I have a shadow network that’s all negative all the time?

SACERDOTE: You could go with Constantly Negative News. But that may be taken.

*      *      *

Freakonomics Radio is produced by Stitcher and Dubner Productions. This episode was produced by Zack Lapinski. Our staff also includes Alison Craiglow, Greg Rippin, Joel Meyer, Tricia Bobeda, Mary Diduch, Brent Katz, Emma Tyrrell, Lyric Bowditch, Jasmin Klinger, Eleanor Osborne, Ryan Kelley, and Jacob Clemente. Our theme song is “Mr. Fortune,” by the Hitchhikers; the rest of the music this week was composed by Luis Guerra. You can follow Freakonomics Radio on Apple Podcasts, Spotify, Stitcher, or wherever you get your podcasts.

Here’s where you can learn more about the people and ideas in this episode:

SOURCES

RESOURCES

“More Than 90 Local Newsrooms Closed During the Coronavirus Pandemic,” by Kristen Hare (Poynter, 2021).
“Why Is All COVID-19 News Bad News?” by Bruce Sacerdote, Ranjan Sehgal, and Molly Cook (Working paper, 2021).
“Out-Group Animosity Drives Engagement on Social Media,” by Steve Rathje, Jay J. Van Bavel, and Sander van der Linden (Proceedings of the National Academy of Sciences, 2021).
“Coping with Stress,” by the Centers for Disease Control and Prevention (2021).
“Newspapers Fact Sheet,” by the Pew Research Center (2021).
“Cable News Fact Sheet,” by the Pew Research Center (2021).
“NBC: Tokyo Olympics May Be Most Profitable Ever Despite Coronavirus Pandemic,” by Gerry Smith and Bloomberg (Fortune, 2021).
“GOP Pushes Bills to Allow Social Media ‘Censorship’ Lawsuits,’” by Anthony Izaguirre (AP News, 2021).
“Silicon Valley Braces for Tougher Regulation in Biden’s New Washington,” by Tony Romm and Elizabeth Dwoskin (The Washington Post, 2021).
In the Same Breath, by Nanfu Wang (2021).
“China Is Now Sending Twitter Users to Prison for Posts Most Chinese Can’t See,” by Chun Han Wong (The Wall Street Journal, 2021).
“More Than Eight-in-Ten Americans Get News From Digital Devices,” by Elisa Shearer (Pew Research Center, 2021).
“Study Finds Around One-Third of Americans Regularly Get Their News From Facebook,” by Sarah Perez (TechCrunch, 2021).
“Ideology, Not Affect: What Americans Want from Political Representation,” by Mia Costa (American Journal of Political Science, 2020).
“Yellow Journalism: The ‘Fake News’ of the 19th Century,” (The Public Domain Review, 2020).
“Revenue of Selected Social Media Companies From 2014 to 2019,” by Statista Research Department (2020).
“Sentiment Mining 500 Years Of History: Is The World Really Darkening?” by Kalev Leetaru (Forbes, 2019).
“The Welfare Effects of Social Media,” by Hunt Allcott, Luca Braghieri, Sarah Eichmeyer, and Matthew Gentzkow (American Economic Review, 2019).
“More Than 1 in 5 U.S. Papers Has Closed. This Is the Result.” by Lara Takenaga (The New York Times, 2019).
“What Makes Online Content Viral?” by Jonah Berger and Katherine L. Milkman (Journal of Marketing Research, 2012).

EXTRAS