As COVID-19 cases continue to rise across the country, Americans remain widely divided in their compliance with recommended public health protocols. New research co-authored by Marketing Science Institute Academic Fellow Jean-Pierre Dubé finds that Fox News viewership is one reason why. Dubé, a marketing professor at the University of Chicago’s Booth School of Business and research associate at the National Bureau of Economic Research, co-authored “The Persuasive Effect of Fox News: Non-Compliance with Social Distancing During the Covid-19 Pandemic,” with Andrey Simonov, a professor of marketing at Columbia Business School, Szymon Sacher, a Ph.D. student in economics at Columbia, and Shirsho Biswas, a recent graduate of the doctoral program at the University of Chicago who is now an assistant professor of marketing at the University of Washington’s Foster School of Business.

Shortly before the 2020 general election, as news of President Donald Trump’s diagnosis with COVID-19 was making headlines across the television and political dial, Dubé and Simonov spoke with Marketing Science Institute Executive Director Barbara Kahn, a marketing professor at The Wharton School of the University of Pennsylvania, to discuss their findings and the potentially wide-ranging implications of those insights.

Listen to the podcast using the player above. An edited transcript of the conversation follows. 

Barbara Kahn: As marketing experts, how did you become interested in this subject?

Jean-Pierre Dubé: The basic idea for this research was [Simonov’s]. He came up with the original idea, but I’ll still just take a quick shot at the answer: The media and communications are a fundamental part of marketing. Andrey was my graduate student when he was in our doctoral program at Booth, and this was a big part of his dissertation. I owe it to Andrey for piquing my interest in communication in terms of media as opposed to just advertising.

We have a real interest in the media, and we have real questions in marketing not only about whether or not we can persuade people to buy stuff, but if more broadly communications and media can change people’s opinions, their beliefs, and their behaviors over things that aren’t just mundane, like what do I buy at the grocery store, but life decisions that can have real consequences for your health and your well-being.

Kahn: Andrey, what drew you to the topic, and why did you choose marketing? And how did all of that go together?

Andrey Simonov: First, this particular work … was not always in my dissertation. My dissertation focused on news markets in general, particularly in Russia. But this naturally fits into a broader scheme of how we think about persuasion in marketing. [Dubé] and I were working, using very similar identification work, or a way to estimate this effect, in the context of persuasion in advertising when COVID-19 started.

A natural extension of thinking about persuasion in marketing, let’s say how advertisers persuade you to buy something, is about how the media persuades people to change their behavior: How it changes beliefs, and then how people [behave] in this case in the quite high-stakes environment where the way people will decide to comply with social distancing can matter a lot. In this way, using the same tools we use in marketing or in broader social sciences, it was a natural question to think about … with other work that we were doing.

Kahn: What were the key takeaways of this research?

Dubé: I would describe the main finding as follows: If we look across geographic regions, places where people tend to watch more Fox News tend to be associated with less compliance with social distancing. People literally leave their homes more in places where there’s more Fox News viewership. They are more likely to go to work even at the height of the crisis — I’m talking not just in March, but in April, for example, during a time when the White House had more or less very unequivocally said, “We totally align with the CDC and their lockdown guidelines.”

A big part of this research was establishing that the association between Fox News viewership and not complying with social distancing was a causal effect, and that Fox News viewership seems to have played a role in causing people not to comply.

Kahn: Can you explain why showing that causal effect is so important, and how that is different from other kinds of approaches, which may show more of a correlational effect?

Simonov: The main challenge — and that is where we started from — is that folks on the [political] right in the U.S. during the early stages of pandemic were not social distancing as much as folks on the [political] left. And if you think about it, we wouldn’t really expect that. There’s no reason for us to expect that Republicans would be less likely to comply.

One theory people had was the idea of messaging. That is where it’s always hard to separate out, as we discussed, because people who watch, let’s say Fox News, which will have very different messaging, will also be people who might choose to social distance for other reasons. That is why it is important. We were trying to find the right natural experiment in the data to be able to get some incremental additional viewership of Fox News and of other news channels for comparison on people across the U.S., and see if it affects the degree to which people comply with social distancing.

Kahn: What was that data? Did you look at other news sources besides Fox News?

Dubé: We looked at the two main competitors, MSNBC and CNN. We had a strategy for trying to isolate this so-called causal effect. We applied it to Fox News, and also to CNN and to MSNBC.

Earlier, Andrey mentioned that we were doing some parallel work on advertising. What we learned was that historically, the way that TV channels got to find the position on their cable box when they entered a geographic market is kind of random. I mean, it’s not that a cable channel can go and say, “I want position 20, or I want position 30.” There’s some randomness in the assignment of these channels.

We also have data now for Nielsen that shows that being in an inferior position, being further away from zero, makes it less likely that you’re going to get viewed. There’s almost a linear decline — the further away you are from zero, the less viewership you get. This is the essence of our so-called experiment: We look at wiggles across markets and what position a channel gets assigned, and we look at the effect of those channel positions on viewership. So, in a sentence, instead of correlating Fox News viewership with social distancing compliance, we instead look at the correlation between the additional viewers you get on Fox News because of your favorable channel position, and we correlate that with compliance. As long as we are convinced that the channel assignment is close to random — and we spent a lot of time defending that and exploring that — … it allows us to see the effect of viewership on compliance.

Kahn: Are there other factors that might influence compliance? How do those factors interact with Fox News viewership?

Simonov: There are two parts to this question. First, there was a big part about those other factors that might correlate with exposure to Fox News, but which we can try to rule out because of the channel position dispute that Jean-Pierre described. Those are all the things we try not to look at, and try to get at the effect of additional Fox viewership or of MSNBC viewership.

The second part of your question is, how can we interpret this? That is when it becomes interesting. One theory you might have is that it’s the effect of the current messaging of Fox News versus MSNBC because Fox News … really was different from other large media in terms of how it covered the pandemic and social distancing guidelines. That is one potential story for the effect.

Another reason can be maybe it’s just that Fox News for the last four years was telling us that experts in mainstream media are not really as trustworthy. Maybe it’s not the current messaging, but the accumulated effect of Fox News telling viewers not to trust experts all the time. Or maybe it’s them translating how politicians on the right are covering this. These are all potential stories which might drive this effect. That is why it is hard for us to separate out exactly which one it is.

Kahn: Does it help to know how often Fox News coverage diverges from the opinions of, say, President Trump, in order to separate out some of these effects?

Dubé: That’s a great question, and that was an important detail in the study. A natural question someone might have is: Is this really a Fox News effect, or is it just that Fox News has created a lot of diehard Republicans, and what we’re really documenting is not just social distancing but how you show affiliation or allegiance with your Republican beliefs. And so, what is important in the study is that on March 13, the White House did announce a national emergency and Donald Trump spoke out very strongly about the dangers [of the coronavirus]. He did say that he recommended lockdown, and he supported the CDC. Through most of the latter half of March and most of April, the White House more or less rallied behind the expert opinion that people should lock down.

Fox, however, continued in several of their broadcasts, not all, to downplay the risks. Laura Ingraham talked about “so-called experts,” she accused the Democrats of the weaponization of fear, that this was a hoax, that expert models had failed. When you look at our effects — even when the White House was telling people to stay at home and be safe — we still see the Fox News effect going full force.

What I think was helpful for us was that prior to March 13, you might say the reason the Fox News effect was kicking in was because it was just echoing Donald Trump’s statements from the White House. But the fact was that after he had stopped making those statements, the Fox News effect persists. That makes us more comfortable that this was about the kind of information people were getting on the media.

I want to say one thing about the question you asked earlier. I think it’s really hard to associate these effects with one specific broadcast or one specific message. It seems like this is an accumulation of information and content that Fox News has been propagating throughout the crisis, and even before the crisis. This is about Fox News’s constant questioning of the reliability of institutions, including scientific institutions.

Kahn: What’s next for this research?

Simonov: One way to frame this was that our research was about the effect of media on people’s behavior in a fairly high-stakes environment that we don’t see often. In this way, even though we dealt with COVID-19 and the pandemic, it’s also a good case study for something we care about outside of this study.

There was a paper on how the media affects people deciding to evacuate from hurricanes, which would be close to our context. The broad theme we’re trying to explore is how it’s the media and not really experts that provide information to the public in the U.S. So, within the COVID-19 pandemic, you can expect that the big next decision will be vaccination compliance. If vaccinations also become a political issue that is covered the same way in the media, you can expect similar effects.

More broadly, like any important effect, [this could apply in any] context where we want to listen to experts, but the media might cover it differently. That is the important part we want to continue to look at.

Kahn: Can you generalize some of these effects to the persuasiveness of advertising?

Dubé: The first answer is more of a methodological one. That is, we’ve now got proof of concept that this method works for establishing a causal effect, in this case of watching a TV show, a news show, on your behavior. This makes me more optimistic that if there is an effect of advertising on shopping, that the same method would help us establish that viewership of a commercial has a causal effect on a person’s behavior when it comes to shopping. The second thing is that we found this persuasive effect of media, and that makes me more confident that advertising messaging in general might be persuasive. If news media can be persuasive, then other communications might be as well.

Kahn: Andrey, do you have something to add to that idea, the connection to advertising persuasiveness?

Simonov: I second the same sentiment. The only thing I would say is that from the literature on both — with separate research on the persuasiveness of advertising and the news media — it looks like in general that the persuasive effect of the news media is much higher than that of advertising.

Usually in the studies it is considered to be much harder to pin down an ad effect. That makes sense because we see a lot of ads, and people don’t tune into Fox News because they want to see the ads. They tune in because they want to see the news. So, I agree completely, it’s a great proof of concept, and we definitely understand that the media has persuasive power, but I wouldn’t extrapolate from that the effect on ads. That’s where I would be careful.

Kahn: There is a famous model in marketing called the Persuasion Knowledge Model, which assumes that when people are exposed to persuasive messages, they will carry out strategies designed to defend against that message. In other words, people will not be as influenced by advertising because their guard is up, whereas with news media they expect they are being told the truth. Is that what you’re saying?

Simonov: Right. And there are models of advertising … where you can also think if I want to see the ad, I expect it to be much more persuasive, and I’d expect to keep looking at the ad. In contrast, if I want to skip the ad, I expect it to be less effective.

Kahn: The general idea that information can still persuade and change opinion and behavior is important. How to make your advertising persuasive is a secondary question. One last question: You looked at traditional news media. Are you looking at all at social media? Are there any feelings about how that may or may not differ? What are your general issues or interest in this area?

Dubé: We deliberately focused on broadcast news because there’s evidence that suggests that almost 50% of Americans rely on televised news as their primary source of information on current events. Fox News is by far the most highly watched channel. I was shocked when I saw our Nielsen data, that Fox News has more viewership than CNN and MSNBC combined.

In the U.S. — and I think this is probably true in other Western countries — trust in institutions has been in decline, and trust in science and scientific experts have been in decline. For example, I saw one statistic that 49% of a random sample of Americans believe that human behavior causes global warming, even though 84% of the scientific community is in alignment [with that view]. So, there is clearly a disjoint between what scientists think and what people think.

I have also had lots of people say, “Well, what if social distancing really didn’t help? Maybe Fox News was doing a good thing.” I say, “That’s missing the point.” I don’t care at this point whether or not you should socially distance. I’d like to think that if people had a doubt about the CDC’s recommendation, they wouldn’t be going to Fox News for their scientific second opinion, the same way that if I get a bad diagnosis from my doctor, I’m not going to turn to CNN as my second opinion.