Does misinformation on Facebook and Instagram change what you believe?

For years, social media companies have faced pressure from users and government watchdogs to stop misinformation from distorting elections, deepening political divisions and undermining Americans’ trust in institutions.

But what if dramatically reducing people’s exposure to unreliable sources that frequently spread misinformation doesn’t actually change what they believe?

UC Berkeley Political Science Ph.D. student Olivier Bergeron-Boutin tested that question on Facebook and Instagram during the 2020 presidential election. In a new paper published in Science Advances, titled “Untrustworthy sources on Facebook and Instagram in 2020(link is external),” Bergeron-Boutin and his coauthors report findings from a large-scale study in which an independent group of academics partnered with researchers at Meta. 

The researchers conducted an experiment among consenting Facebook and Instagram users that reduced social media feed exposure to content from sources that repeatedly spread misinformation. 

The experiment was successful in changing what people saw: participants assigned to the intervention were exposed to roughly 70% less content from untrustworthy sources. However, their political beliefs and attitudes did not measurably change, even among those who were previously exposed to the most content from those sources. 

“The intervention was successful in the sense that it reduced what it was supposed to reduce, in this case, exposure to information from untrustworthy sources,” Bergeron-Boutin said. “But the resulting changes in attitudes were much smaller than some would have expected.”

These findings complicate the assumption that exposure to content from dubious sources that spread misinformation on social media powerfully shapes people’s beliefs and attitudes. The study instead suggests that while social media companies can change what people see, combating misinformation and political polarization may require confronting deeper forces that shape what people believe.

Feeds changed, beliefs didn’t

The researchers also found that exposure to untrustworthy sources was not evenly spread across Facebook and Instagram users. Most people on both platforms saw relatively little content from these sources. Instead, a relatively small group of users was responsible for most of the exposure — 23% of Facebook users and just 11% of Instagram users — accounted for roughly 80% of all exposure to information from untrustworthy sources. 

“Exposure to content from untrustworthy sources on social media is less common and more concentrated than people typically think,” said Brendan Nyhan, the James O. Freedman Presidential Professor in the Department of Government at Dartmouth College and a co-author of the study. “In addition, reducing exposure to content from those sources did not have the effects that many speculated it would have on people’s beliefs, attitudes and behaviors.”

UC Berkeley computational folklorist Tim Tangherlini, who studies rumors and conspiracy theories on social networks but was not involved in the Facebook/Instagram study, said that concentration reflects something researchers have long observed about how information moves through communities.

“Stories don’t spread evenly through a population,” Tangherlini said, noting the timeliness and importance of the study. “They flow on and across social networks predisposed to receive them, retell them and refine them — basically like an echo chamber.”

What happens when you take people out of the echo chamber?  

Researchers then tested what happened when people saw far less of that content. For three months surrounding the 2020 election, they reduced participants’ exposure to untrustworthy sources by roughly 70%.

Yet people’s political beliefs and attitudes did not measurably change.

Tangherlini said the finding makes sense when political beliefs are understood as something shaped not just by the information people encounter online, but also by the communities and social networks around them.

“A platform intervention changes what flows into the network,” Tangherlini said, “but it doesn’t change the underlying social network, or the beliefs shared in that network.”

He compared changing deeply rooted beliefs to steering a large oil tanker: small corrections can change its direction, but significant changes may take time.

A debate that has only grown

Since 2020, social media platforms have changed substantially. Meta ended its U.S. third-party fact-checking program in 2025, eliminating the system of fact-checking “strikes” that made the study’s source-level intervention possible. The researchers caution that their 2020 findings, therefore, cannot tell us what misinformation exposure looks like on Facebook and Instagram today.

Nyhan added that the findings point to the need to focus on those most heavily exposed rather than treating misinformation as a problem that affects everyone equally.

“I hope social media companies and policymakers will recognize that exposure to content from untrustworthy sources is heavily concentrated and the harms from it are likely to be similarly concentrated,” Nyhan said. “We should worry less about the average person (who is typically uninterested in politics) and more about potential effects on people who consume large volumes of dubious content online and what it might do to them or inspire them to do.”

Bergeron-Boutin said the study’s results should not discount the problems misinformation poses.

“We shouldn’t interpret these results to mean that misinformation is no big deal,” he said. “But there’s no quick fix. Anyone who proposes a simple solution to misinformation or political polarization should be viewed skeptically. Tinkering at the margins can help, but ultimately, institutional features of American democracy provide fertile ground for these issues to rise.”

More questions, less transparency

The study also highlights another challenge: researchers need access to the platforms themselves to understand how social media affects society, but that access is becoming harder to obtain.

“This study presents a unique opportunity because social media platform data is rarely accessible to outside researchers,” Bergeron-Boutin said. “A study like this is becoming less and less likely to happen again, because since 2020, these platforms have become less transparent.”

For Nyhan, that makes accurately measuring the harms caused by misinformation — and determining how society should respond — one of the most important challenges ahead.

“We need to learn how to better measure the potential harms from dubious content online and address them without compromising the values and freedoms that we hold dear,” he said. 

As platforms become less transparent, answering those questions may become increasingly difficult. The study shows that social media companies can change what millions of people see. Understanding how that shapes what they believe is a much harder question.

The study’s other lead co-authors are Jaime Settle, department chair and professor of government at The College of William & Mary; Emily Thorson, associate professor of political science at Syracuse University; and Magdalena Wojcieszak, professor of communication at UC Davis.

---

---