IE 11 is not supported. For an optimal experience visit our site on another browser.

Do Facebook and Instagram influence your politics? 35,000 took a break to find out

A study published Monday said staying off social media apps had nearly zero effect on how people viewed Biden and Trump in 2020 but did have other effects.
The Facebook app logo is displayed on a smartphone.
Researchers wanted to understand how not using Facebook and Instagram would affect individual users' politics.Thomas Trutschel / picture-alliance / dpa / via AP file

Facebook and Instagram may not have quite the impact on people’s politics as some critics have feared. 

A study published Monday by researchers at Stanford University, Meta and other institutions found that a subset of 35,000 people who took breaks from Facebook and Instagram for six weeks before the 2020 presidential election didn’t significantly change their politics by the time the election came around. 

The researchers found that staying off the two social media apps, both part of parent company Meta, had nearly zero effect — at least in the short term — on how people viewed the candidates, whether they turned out to vote and how they perceived the legitimacy of the election. 

But the researchers also found that staying off Facebook, in particular, was a double-edged sword for understanding the world: Those who deactivated the app appear to be less likely to believe misinformation about the election but also less likely to be knowledgeable about general news. 

The research, published in the Proceedings of the National Academy of Sciences, is part of a broader study trying to understand how social media affects democracy. Under fire for years for upending how campaigns are run and how voters consume information, Meta formed a partnership with academics to grant them access to company-held data. University professors retain control over what they report. In the latest study, the authors said that Meta paid for the costs of the research but that it didn’t pay the researchers or their institutions.

Meta said in a statement Monday: “These findings are consistent with previous publications in this study in showing little impact on key political attitudes, beliefs or behaviors.” An earlier set of findings published in July concluded that Facebook had “significant ideological segregation” and that conservative sources dominated its news ecosystem. 

Academic researchers and social media users alike have been eager to understand, for example, whether so-called filter bubbles exist and if so what the practical effects are. Social media has been repeatedly blamed for “ruining political discourse,” spreading falsehoods and making polarization worse, and research into such questions is relatively young. 

A co-author of the latest study, Matthew Gentzkow, a Stanford economist, said one of the study’s findings was how little impact Facebook and Instagram seemed to have on polarization and divisiveness. 

“If we’re worried about those things, trying to control what people see on social media and whether they’re on social media may not be the most important lever,” he said in an interview. 

On issues such as immigration, mask requirements and policing, those who deactivated Facebook and Instagram remained polarized. 

But Gentzkow also said the study isn’t nearly the final word on the subject, because it was limited to the impact of Facebook and Instagram in a relatively narrow six-week time frame. 

“This study cannot say one way or the other — in a decadelong sense — whether social media is causing polarization or not,” he said. 

Researchers did find a small, unverified impact of Facebook use favoring Donald Trump, in which people who deactivated from Facebook were slightly more likely than others to vote for Joe Biden — the equivalent of 1.3% of Trump voters’ swinging to Biden. That could be because the Trump campaign was using Facebook more effectively or possibly because of other factors, the researchers wrote. Either way, they wrote, the difference “applies to the specific population that selected into our experiment” and “cannot be extrapolated to the broader population without strong assumptions.” 

The study bills itself as the “largest-scale evidence available to date on the effect of Facebook and Instagram access on political knowledge, attitudes, and behavior in a presidential election season.” It lists 32 co-authors from 14 institutions. 

Researchers didn’t look at the role of Facebook and Instagram after Election Day, when some supporters of Trump used them to promote allegations about election fraud and the company battled to take the allegations down

Meta worked with the researchers to find participants, placing invitations in people’s feeds in August and September 2020. And participation came with a financial incentive: $25 apiece for people who agreed to deactivate for one week and $150 for people to deactivate for six weeks. 

It’s not the first study to wonder what life is like for people not on social media. A study published in 2018, amid an online boycott called #DeleteFacebook, said that “the average Facebook user would require more than $1,000 to deactivate their account for one year.”

And in a study from 2019, Gentzkow and other researchers found that deactivating Facebook for four weeks before the 2018 midterm elections had a variety of real-world implications, including increased self-reported happiness, reduced factual news knowledge and reduced political polarization. Gentzkow said more research would be needed to explore why that study showed reduced polarization but the most recent one did not.