On media diets, partisan filtering, and getting uncomfortable

April 19, 2022
AP Photo/Gerald Herbert

In late 2020, David Broockman and Josh Kalla, both political science professors, conducted what they later termed “a field experiment with Fox News viewers.” Broockman and Kalla gathered an initial pool of registered voters who regularly watched Fox News (between 500 and 14,400 minutes per month) and rarely watched CNN (less than 30 minutes per month). They winnowed that pool to roughly 800 people who, taken together, self-reported watching 14 hours of Fox News each week. Of those people, 90 percent voted in the 2016 general election, and 92 percent self-identified as Republican. More than 300 were assigned to a “treatment group,” whose members were paid $15 an hour to watch CNN instead of Fox News and were quizzed regularly on content. Members of the treatment group watched nearly six hours of CNN per week, on average, during the experiment. The experiment ran for more than three weeks in September 2020; during that time, the networks “covered dramatically different topics,” with CNN focusing on “covid and election integrity” and Fox News on “issues related to race and protests.”

Recently, Broockman and Kalla published their findings. Broockman took to Twitter to summarize them, writing at the outset, “Our results should worry you.” 

In surveys with their treatment group, Broockman and Kalla “found large effects of watching CNN instead of Fox News on participants’ factual perceptions of current events and knowledge about the 2020 presidential candidates’ positions.” Summarizing the findings online, Broockman wrote that those Fox News viewers who switched to CNN were “less likely to agree” that Fox News would discuss negative actions by Donald Trump, and liked Trump less than those who kept their Fox News habit; they were also “more supportive” of vote-by-mail initiatives and “more negative in their evaluations of Donald Trump and Republican politicians.” Not everything changed: Broockman and Kalla noted that new CNN viewers did not show shifts in partisan identification, nor did they change their attitudes “around race, climate change, and policing.”

The study kicked off a round of coverage by media watchers. “What if Fox News viewers watched CNN instead?” ran a headline at Bloomberg, where Matt Yglesias detailed the range of changes. The new CNN viewers “were 10 points less likely to believe that supporters of then-candidate Joe Biden were happy when police officers get shot, 11 points less likely to say it’s more important for the president to focus on containing violent protesters than on the coronavirus, and 13 points less likely to agree that if Biden were elected, ‘we’ll see many more police get shot by Black Lives Matter activists,’ ” Yglesias wrote. “These are reasonably large changes from a one-month experiment. And they occurred despite the long-term effort of then-president Donald Trump to discredit CNN and other mainstream media outlets.”

Elsewhere, Sree Sreenivasan, a journalist and professor, noted the “alternate reality” of Fox News viewers and called for those viewers to “diversify your media intake, and—for lack of a better phrase—do the work” of finding coverage from more than one source. On CNN, Brian Stelter spoke with Broockman and Kalla, above a chyron that read, “What happens to Fox viewers when they change channels?” Former president Barack Obama raised the study during an interview with The Atlantic’s Jeffrey Goldberg. “I think we underestimate the degree of pliability in our opinions and our views and what that means,” Obama said. “I take that as hopeful.”

Such optimism characterized many of the pieces covering Broockman and Kalla’s study. After all, if diehard conservatives are more likely to believe that Donald Trump could have handled covid better, as some did after watching CNN, then perhaps everyone has the capacity to believe information that runs counter to their personal opinions and preferred news sources. 

Sign up for CJR's daily email

In their report, however, Broockman and Kalla noted that “partisan coverage filtering”—a term they use to characterize selective reporting that “leads viewers to learn a biased set of facts”—breaks both ways. Kalla told Stelter during their segment that “CNN engages in this partisan coverage filtering as well,” noting that Fox News covered the Abraham Accords “about fifteen times more than CNN did” and that the issue at hand was “about the media writ large.” 

“I think you’re engaging in some both-sides-ism there, Josh,” Stelter replied. Kalla responded that, to try and get beyond partisan coverage filtering, “we need viewers to see all types of information. Unfortunately, what we find in this study is that viewers don’t want to engage in watching all sides.”

Though Fox News viewers gained new beliefs when exposed to new material, they retreated to the right-wing outlet once the study was over. “Even though we try to incentivize viewers to watch both Fox and CNN, they don’t want to engage in that hard work,” Kalla said on Stelter’s show. “They want to really just watch the side that makes them feel good. And this is why the media has such an important responsibility to cover both sides, to hold both parties accountable.” 

I wrote about this very problem after resigning from MSNBC in July 2020. (“Our viewers don’t really consider us the news,” a senior producer had previously told me. “They come to us for comfort.”) That tendency by audiences—to selectively expose themselves to partisan news—risks creating a damaging feedback loop in which producers prioritize audience comfort and ratings at the expense of more expansive coverage. 

Broockman, who spoke with me by phone, returned several times to “democratic accountability”—the idea that media outlets play a central role in helping voters hold elected officials accountable. (“Hiding information from voters can undermine their ability to hold their elected officials accountable,” Broockman and Kalla wrote near the end of their study. “Our evidence indicates that partisan media do exactly this, with manifold consequences for their beliefs and attitudes.”) Even if journalists or networks favor a certain party, Broockman told me, they should always remember that they have a lot of power to shape the perspectives of those politicians and their constituents. He noted that many people who voted for Trump in 2016 also believed he had done a good job by 2020, a perspective he attributed in part to favorable coverage from Fox News. But he said the same dynamic could extend to, say, MSNBC’s coverage of House Speaker Nancy Pelosi. Critical coverage of a favored candidate might inspire critical feelings among that candidate’s supporters, prompting the candidate in question to do better, Broockman explained. However, “if you give them a blank check, that won’t make politicians work hard.”

None of this is to draw an equivalence between the partisan coverage filtering of Fox News and that of CNN or MSNBC. My conversation with Broockman called to mind a November report from the Kaiser Family Foundation’s covid-19 Vaccine Monitor, which found a correlation between “trusted news sources” and “belief in covid-19 misinformation” and stated that 36 percent of respondents who trust Fox News believed or were unsure about four or more false statements, of a total of eight. According to the same report, “at least a third of those who trust information from CNN, MSNBC, network news, NPR, and local television news do not believe any of the eight false statements.” (The report did not evaluate whether viewers were “exposed to misinformation from those news sources” or were “pre-disposed to believe certain types of misinformation for other reasons.”)

Let’s make sure the lessons from this important study reach their broadest possible audience. Yes, it confirmed that Fox News viewers may be misinformed. But, on a far more hopeful note, it suggested that those viewers, and others, have the ability to absorb information necessary to hold democratic institutions accountable—so long as they are exposed to it. Perhaps outlets could consider reporting more critically on the politicians who actually represent those partisan audiences. In turn, maybe audiences could encourage such accountability reporting by sticking with those news stories they might find uncomfortable.

Ariana Pekary is the CJR public editor for CNN. She was an award-winning public radio and MSNBC journalist for two decades. Now she focuses on the systemic flaws of commercial broadcast news. She can be contacted at publiceditors@cjr.org or on Twitter @arianapekary.