On January 6, Facebook removed two of Trump’s posts (for violating its Community Standards on Dangerous Individuals and Organizations) as his supporters stormed the Capitol. Facebook also suspended Trump’s ability to post on Facebook and Instagram for 24 hours, extending the suspension the next morning “indefinitely and for at least the next two weeks until the peaceful transition of power is complete.”

This is not the first time that Facebook has had to take action against Trump’s posts. Facebook removed at least one post with COVID-19 misinformation as its rules prohibit content with “the potential to contribute to real-world harm,” and it applied clarifying labels on posts containing false claims about the presidential election. In fact, Media Matters has identified roughly 500 posts from Trump between January 1, 2020, and January 6, 2021, that Facebook labeled with authoritative information in an attempt to stop his false claims from spreading. As many have noted, these half-measures are ineffective and characteristic of Facebook's handling of many topics. Even Facebook’s own internal data reportedly indicates that the labels failed to slow the spread of Trump’s false claims. 

Facebook’s oversight board is now tasked with determining whether the site correctly decided to prohibit Trump from posting content on its platforms. The board, which was created “to offer a critical independent check on Facebook’s approach to moderating some of the most significant content issues,” is expected to announce a decision on Trump’s accounts by April. Notably, the oversight board has been asked to consider only two of Trump’s posts in its deliberations — the two removed by Facebook as rioters stormed the Capitol. By looking at only these two posts, and potentially ignoring the rest of Trump’s posts, the board risks missing the bigger picture -- that these posts were part of a much broader pattern, as Trump repeatedly used his Facebook page to attack others and spread harmful misinformation.

Media Matters analyzed Trump’s 6,081 Facebook posts that were created between January 1, 2020, and January 6, 2021, when Facebook suspended his account. We found 1,443 posts that contained COVID-19 misinformation, election misinformation, or harmful rhetoric attacking his critics as extreme or fake. These posts, which accounted for roughly 24% of Trump’s posts since January 1, 2020, represent a pattern of behavior by the then-president to bolster his own influence at the risk of undermining public trust of institutions. 

Other Facebook accounts have been permanently banned or removed for far less. Facebook has banned far-right figures, such as Alex Jones, saying the platform always bans accounts “that promote or engage in violence and hate, regardless of ideology.” Similarly, it banned pages affiliated with Roger Stone and Steve Bannon for spreading misinformation.

Facebook states that “multiple violations may result in your Page being unpublished, having limits placed on it or other restrictions,” but it does not specify how many violations will result in a temporary block or permanent suspension of an account. Reporting has previously indicated that when a page administrator receives five “strikes” within 90 days, the page will be deleted. Posts shared by a page that incite violence or contain misleading content are considered violations.

Despite these standards, there are multiple posts on Trump’s page that seemingly incite violence; altogether, they have earned over 2.2 million interactions.

Facebook also labeled hundreds of Trump’s posts with authoritative information about the election, noting, “The US has laws, procedures, and established institutions to ensure the integrity of our elections” or “Election officials follow strict rules and have found no evidence of widespread fraud.” Despite these labels on Trump’s posts, Facebook did not place any restrictions on Trump’s account until after these false claims inspired violence at the Capitol.

Trump’s 1,443 posts with misinformation and extreme rhetoric attacking his critics also earned high engagement, with over 331.6 million interactions. Notably, these posts earned roughly 230,000 average interactions per post, which is more than the roughly 152,000 average interactions per post for all of Trump’s posts for 2020 and the beginning of 2021.

Facebook removed at least one post with COVID-19 misinformation as its rules prohibit content with “the potential to contribute to real-world harm,” but Media Matters has identified at least 516 posts Trump created between January 1, 2020, and January 6, 2021, that contained misinformation about COVID-19. These posts accounted for 8% of Trump’s posts since January 1, 2020, and earned over 83.9 million interactions or an average of 162,000 interactions per post.

In posts about COVID-19, Trump frequently minimized the threat of the virus, falsely equating it with the seasonal flu, falsely suggesting that cases and deaths are exaggerated, falsely claiming that the mortality rate in the U.S. is one of the lowest in the world, and promoting hydroxychloroquine as a treatment for COVID-19 despite no substantial evidence proving its effectiveness.

“Testing”: Trump often claimed that higher numbers of COVID-19 cases in the U.S. as compared to other countries were a result of more testing here, which has been debunked. We found that Trump mentioned “testing” at least 55 times, earning 7.6 million interactions on posts with the phrase or an average of 139,000 interactions per post. One post about testing earned over 573,000 interactions:

“China Virus” or “Chinese Virus”: Trump and his allies in right-wing media have repeatedly used racist and xenophobic language when discussing COVID-19, calling the virus the “China Virus” or “Chinese Virus,” even though medical experts have warned against using such rhetoric. He also used this language on Facebook at least 50 times between January 1, 2020, and January 6, 2021, earning roughly 11.7 million interactions or an average of 244,000 interactions per post with these phrases. One post with the racist language earned nearly 1 million interactions:

Other language Trump included in posts with COVID-19 misinformation included “flu,” “death rate,” and “hydroxychloroquine”:

Total mentions of — and engagement with — keywords or phrases related to COVID-19 misinformation in Trump’s 2020, 2021 Facebook posts

Facebook interactions include reactions (like, love, wow, haha, sad, angry), comments, and shares.

From a study of the 6,081 Facebook posts former President Donald Trump created between January 1, 2020, and when his account was suspended on January 6, 2021.

Facebook struggled to stop election misinformation as Trump and his allies repeatedly attempted to undermine the integrity and results of the 2020 presidential election. In fact, we found at least 363 posts from Trump between January 1, 2020, and January 6, 2021, that contained election misinformation. These posts accounted for 6% of Trump’s posts since January 1, 2020, and earned over 149.5 million interactions or 412,000 average interactions per post.

“Stop the Steal”: Notably, Trump mentioned “Stop the Steal” at least 10 times, with some of these posts promoting “Stop the Steal” events. Posts with “Stop the Steal” or “stopthesteal” earned roughly 4.4 million interactions or an average of 486,000 interactions per post, which is more than 3 times the average interaction per post on all of Trump’s posts since January 1, 2020. After pro-Trump supporters breached the Capitol on January 6, Facebook implemented a narrow ban on “Stop the Steal” content, but Trump’s posts mentioning “Stop the Steal” are still on Facebook.

Seven of these posts were actually created in the five days leading up to the riot at the Capitol:

“Dominion”: Trump specifically mentioned “dominion,” referring to a voting machine company that has been a frequent target for right-wing media, at least 15 times since January 1, 2020, earning over 10 million interactions or an average of 683,000 interactions per post. One post mentioning “dominion” earned over 1.9 million interactions:

Other language that Trump used when discussing the election included “rigged,” “stolen,” and “fraud”:

Total mentions of — and engagement with — keywords or phrases related to election misinformation in Trump’s 2020, 2021 Facebook posts 

Facebook interactions include reactions (like, love, wow, haha, sad, angry), comments, and shares.

From a study of the 6,081 Facebook posts former President Donald Trump created between January 1, 2020, and when his account was suspended on January 6, 2021.

During his presidency, Trump repeatedly used dehumanizing language and vilified his critics -- including the media, Democrats, and even some Republicans -- contributing to increased distrust of institutions and polarization. Trump frequently used Facebook to push this rhetoric to a broader audience, with seemingly little regard for the consequences. In fact, we found at least 683 posts created by Trump between January 1, 2020, and January 6, 2021, that contained such extreme rhetoric, frequently targeting the “fake news” media and “radical left” Democrats. These posts accounted for 11% of Trump’s posts since January 1, 2020, and earned roughly 130 million interactions or 190,000 average interactions per post.

“Fake news”: Trump used the phrase “fake news” at least 217 times since January 1, 2020, earning over 45.5 million interactions or 210,000 average interactions per post. At least 3 of these posts earned over 1 million interactions apiece:

“Radical left”: Trump used the phrase “radical left” at least 145 times since January 1, 2020, earning roughly 26.1 million interactions or an average of 184,000 interactions per post on posts with the phrase. At least three of these posts earned over 1 million interactions, with one even earning nearly 2 million interactions:

“Antifa”: Posts mentioning “antifa” earned nearly double the average interactions per post (300,000) compared to the average for all posts with extreme rhetoric (190,000). Trump mentioned “antifa” at least 23 times since January 2020, earning over 6.6 million interactions on these posts. In one post, which earned over 955,000 interactions, Trump claimed, “The United States of America will be designating ANTIFA as a Terrorist Organization.”

Additional language that Trump used in attacking perceived critics included “far-left,” “corrupt,” and “lamestream media”:

Total mentions of — and engagement with — extreme keywords or phrases used to attack critics in Trump’s 2020, 2021 Facebook posts

Facebook interactions include reactions (like, love, wow, haha, sad, angry), comments, and shares.

From a study of the 6,081 Facebook posts former President Donald Trump created between January 1, 2020, and when his account was suspended on January 6, 2021.

Using Crowdtangle, Media Matters compiled a list of 6,081 posts created by Trump from January 1, 2020, to January 6, 2021, when Facebook suspended his account. We then compiled posts that were related to COVID-19, the presidential election, or Trump’s critics. 

The resulting lists included 607 posts with COVID-19 related keywords, 1,045 posts with keywords related to the election, and 1,597 posts with harmful or dehumanizing rhetoric.

Three researchers independently coded each post to determine whether: 1) Facebook put a label on the post, 2) the post qualified as an incitement of violence, and 3) the post's main narrative was actually about COVID-19, the election, or contained harmful rhetoric directed at his critics.

We reviewed each post individually, and each post was given a final code if two of the three researchers independently awarded it the same code. The posts containing keywords from multiple categories were given multiple codes. 

We searched for the following keywords in the post’s message, description, or image text to determine if it contained COVID-19 misinformation or talking points: “cases,” “China plague,” “China virus,” “China's fault,” “chinavirus,” “chloroquine,” “corona,” “coronavirus,” “COVID-19,” “COVID,” “cure cannot be worse,” “death rate,” “deaths,” “Democrat mayor,” “Democratic governor,” “face mask,” “Fauci,” “fault of China,” “front line workers,” “frontline workers,” “hydroxychloroquine,” “hydroxyzine,” “immune,” “infection rate,” “infection,” “it was China,” “kung flu,” “liberate,” “lockdown,” “lockdowns,” “more testing,” “mortality,” “Operation Warp Speed,” “operation warpspeed,” “Plaquenil,” “shut down,” “shutdown,” “shutdowns,” “stock pile,” “stockpile,” “swine flu,” “test,” “tested,” “testing,” “tests,” “the virus,” “therapeutics,” “vaccine,” “vaccines,” “ventilator,” “ventilators,” “virus,” “W.H.O.,” “World Health Organization,” or “Wuhan”.

We searched for the following keywords in the post’s message, description, or image text to determine if it contained election misinformation or talking points: “biggest scam,” “certification,” “certify,” “certifying,” “chosen electors,” “corrupt,” “corruption,” “count,” “decertification,” “decertify,” “decertifying,” “democratic process,” “Dominion,” “election fraud,” “election victory,” “election,” “electoral college,” “found ballots,” “fraud,” “fraudulently,” “Giuliani,” “Hawley,” “Jan 6,” “Jan 6th,” “Jan. 6,” “Jan. 6th,” “January 6,” “January 6th,” “January sixth,” “landslide,” “protect,” “recount,” “save America,” “signature verification,” “stolen election,” “Stop the Steal,” “StoptheSteal,” “swing states,” “vote dump,” “vote tabulation,” “voter dump,” “voter fraud,” “voter tabulation,” or “surrender caucus.”

We searched for the following keywords in the post’s message, description, or image text to determine if it contained harmful or dehumanizing rhetoric: “agitator,” “agitators,” “anarchist,” “anarchists,” “anarchy,” “anti-police,” “antifa,” “arson,” “arsonist,” “arsonists,” “Black Lives Matter,” “BLM protester,” “BLM protesters,” “BLM protestor,” “BLM protestors,” “BLM,” “burning our cities,” “chaos,” “church on fire,” “crazies,” “crazy,” “crime,” “defund the police,” “Democrat cities,” “Democrat city,” “Democrat run cities,” “Democratic cities,” “Democratic city,” “destroy our country,” “domestic terrorism,” “domestic terrorists,” “extremist,” “extremists,” “fake news,” “far-left,” “flag-burner,” “flag-burning,” “friendly protesters,” “friendly protestors,” “held responsible,” “hold them responsible,” “lamestream media,” “law & order,” “law and order,” “law enforcement,” “lawlessness,” “left-wing,” “liberal cities,” “liberal city,” “liberate,” “loot,” “looter,” “looters,” “looting,” “lowlifes,” “national guard,” “pathetic,” “peaceful protesters,” “peaceful protestors,” “protect,” “puppet,” “puppets,” “radical left,” “radical leftist,” “radical leftists,” “riot,” “rioter,” “rioters,” “scum,” “shooting,” “sleepy Joe,” “terrorist organization,” “terrorists,” “thug,” “thugs,” or “violent groups.”

The final lists included 516 posts with the COVID-19 related keywords, 363 posts with keywords related to the election, and 683 posts with harmful or dehumanizing rhetoric. Overall, 1,443 posts contained COVID-19 keywords, election-related keywords, and/or harmful rhetoric.