Twitter blocked and labeled Donald Trump's tweets on election fraud. They spread anyway.

Jessica Guynn

Twitter blocked and labeled some of Donald Trump's claims of election fraud in the run-up and aftermath of the 2020 presidential election.

The tweets spread on and off Twitter anyway.

That’s according to a new study from New York University researchers published Tuesday in Harvard Kennedy School Misinformation Review and shared exclusively with USA TODAY.

The study is raising questions about the ability of social media companies to halt the flood of falsehoods on mainstream social media platforms during election cycles.

NYU researchers say Trump tweets with fact-check labels spread further on Twitter than those without. And when Twitter blocked engagement with the former president’s tweets, they leaped to Facebook, Instagram and Reddit where they were more popular than tweets that Twitter labeled or did not flag at all.

Do Facebook and Google censor conservatives? Trump, Republicans bet claims they do will rally GOP base in 2022

Banning Trump from Facebook: Why Justice Clarence Thomas and other conservatives say it is First Amendment issue

It's not clear if Twitter intervened on social media posts that were more likely to spread or if it was the intervention itself that gave the tweets a boost, the researchers said.

But they say their study underscores how harmful misinformation can hop from platform to platform with too little coordination among social media companies to curb its spread.

“Misinformation halted on one platform does not halt it on another,” said Megan Brown, research engineer with NYU's Center for Social Media and Politics.

Blocked on Twitter, Trump’s tweets turned up on Facebook in the form of links, quotes and screenshots, where they garnered an average engagement of more than 300,000, said Zeve Sanderson, executive director of the NYU center.

That phenomenon shows that “political actors seeking to advance a narrative online are not limited to working within a single platform,” said Joshua Tucker, co-director of the center. 

“We are in a world where people who are trying to control information environments and who are trying to push political information environments are in a multiplatform world,” Tucker said. “Right now, the only way we have to deal with content is on a platform-by-platform basis.”

President Trump's final Tweet read, “To all of those who have asked, I will not be going to the Inauguration on January 20th.”

In a statement, Twitter said it took a number of steps to limit engagement on tweets that violated its rules.

"As election conversation reached record highs, it was critical that we took swift enforcement action on misleading content that could contribute to offline harm," the company said.

From Oct. 27 to Nov. 11, Twitter labeled some 300,000 tweets as disputed or potentially misleading and saw an estimated 29% decrease in quote tweets. 

"We continue to research, question and alter features that could incentivize or encourage behaviors on Twitter that negatively affect the health of the conversation online or could lead to offline harm," the company said.

Twitter's most significant intervention was permanently banning Trump in the final days of his presidency after the Jan. 6 Capitol attack, a move that raised thorny questions of free speech and censorship on social media. 

At the time, Trump had 88.7 million followers who retweeted him at an astonishing rate, giving him near unprecedented power to shape the national conversation.

After his followers stormed the U.S. Capitol Building to block Congress from certifying Joe Biden’s presidential win, all three of the nation’s top social media platforms – Facebook, Google’s YouTube and Twitter – banned Trump over concerns he would incite more violence.

YouTube said it would lift the suspension after the "risk of violence" decreases. In June, Facebook said the earliest Trump would regain access to his accounts would be 2023. Even if Trump runs for president and wins in 2024, Twitter said it  would not reinstate him.

Trump attacked social media companies for labeling, restricting or removing his posts that spread falsehoods about the outcome of the presidential election.

In July, Trump filed suit against Facebook, Google, Twitter and their CEOs, claiming the companies violated his First Amendment rights.

In a backlash from conservatives, dozens of states are considering legislation that targets how social media platforms regulate speech. One bill passed in Florida but was temporarily blocked by a federal judge.

Another in Texas had the votes it needed in a special session of the Republican-controlled legislature, but it has been in limbo after Democrats left the state for Washington to protest a GOP effort to overhaul the state election system.

NYU researchers say they focused on Trump's tweets “because of evidence that he acted as a central vector for spreading election-relation misinformation.”

They examined tweets from Nov. 1, 2020, through Jan. 8, 2021, that were flagged by Twitter. 

Blocking engagement with Trump’s tweets limited their spread on Twitter but not elsewhere, researchers found. The tweets were posted more often and were more popular on other social media platforms.

When Twitter slapped a warning label on Trump’s tweets, they were more popular than his tweets that had no label, researchers said.

The finding does not necessarily mean that warning labels were ineffective or had the “Streisand effect,” when an attempt to hide or remove information draws even more attention to it, Sanderson said. It may be that the types of tweets that Twitter labeled were also the type that would be more likely to spread.

“In the future, especially with respect to the ongoing pandemic and the 2022 midterms coming up, it will be really important for the platforms to coordinate in some way, if they can, to halt the spread of misinformation,” Brown said.