In general, those who tweeted about politics were also more likely to follow others who share their political views and to feel more harshly toward members of the opposing political party.
They are also overwhelmingly against President Trump.
Twitter and other social media companies tout their ability to expose people to multiple points of view. But it frequently doesn’t work that way, in large part because of the tech companies’ algorithms.
As the United States has become more politically divided, so, too, have social media users. Because of the ability to follow like-minded users and block those who disagree, people increasingly have the opportunity to find themselves in a social media “filter bubble." Those bubbles can be exacerbated by companies’ algorithms, which tend to recommend joining groups and following other users who agree with you.
But an echo chamber of voices that agree with you can prove dangerous, researchers say, especially if infiltrated by agents hoping to influence an election.
When information bounces through the chamber, few users take the time to determine its origin and assess its truthfulness, said David Levine, an associate professor at the Elon University School of Law in Greensboro, N.C., and the founder of the “Hearsay Culture” radio show about modern technology issues.
“It can be quite dangerous if you’re not taking a step back and saying, ‘What do I know about the sources of this information and who or what is behind it?’” he said. “It’s very easy psychologically, especially if you’re coming into it with a particular perspective, to go along with it.”
Twitter declined to comment.
The company’s stock fell 19 percent as the company reported disappointing results stemming from problems with Twitter’s ad targeting technology. Still, Twitter reported strong user growth Thursday, adding 6 million users over the last quarter, for a total of 145 million.
The political dangers presented by social media became clear in the wake of the 2016 U.S. presidential election, when a Russian-backed group called the Internet Research Agency sowed social discord online in an attempt to interfere with the political process. The organization used a variety of digital disinformation tactics — including fake accounts posting about divisive issues — to attack Democratic presidential candidate Hillary Clinton and support Republican candidate Donald Trump, who won the election.
On Monday, Facebook said it removed accounts that appeared to be from the same group that praised Trump and attacked former vice president Joe Biden, a 2020 presidential candidate.
Zuckerberg told lawmakers that Facebook’s defenses against disinformation are more “sophisticated than any other company has at this point, and frankly, governments, too.”
Facebook did not respond to requests for comment.
Pew studied more than 1 million tweets posted between June 2018 and July of this year by 2,427 participants with public Twitter accounts. Among users who tweeted about politics, 72 percent strongly disapprove of Trump, 25 percent strongly approve of him and the others had milder opinions on either side.
Pew found that there is a higher percentage of U.S. adults on Twitter who strongly disapprove of the president than is reflected in the country’s population. The organization also noted that those who are strongly anti-Trump are also more likely to tweet about national politics than other groups on the site.
Conservatives have frequently accused tech giants of exhibiting bias against them. While tech companies say they treat all political content equally, conservatives allege that Silicon Valley’s liberal culture and campaign contributions to Democrats paint a different picture. Trump and other prominent Republicans have repeatedly accused Facebook, Google and Twitter of censoring conservatives or making their content less accessible to the public.