For years, US Republicans have complained that they are mistreated by social media companies, either through subtle algorithms that reduce their reach or through flat-out censorship. But a New York University study concludes the opposite, dispelling the myth that big tech is censoring conservatives.

“After close review of recent Tweets from the @realDonaldTrump account and the context around them — specifically how they are being received and interpreted on and off Twitter — we have permanently suspended the account due to the risk of further incitement of violence,” a Twitter blog post read on the 8th of January 2021.

The decision came on the heels of the 2021 US Capitol insurrection, where nearly 200 officers were injured and 5 people were killed. It was an event many felt was fueled by Donald Trump, who refused to acknowledge Joe Biden as president and spread countless conspiracy theories about the elections, oftentimes on social media. Trump reacted by calling the ban an assault on democracy and Twitter a “Far Left platform”.

While Trump certainly took things to an unprecedented level, his statements were in line with ideas spread by republicans for a number of years: the idea that social media companies are somehow biased pro-left and against them.

‘Conservatives commonly accuse the major social media companies of censoring the political right. In response to Twitter’s decision on January 8, 2021, to exclude him from the platform, then-President Donald Trump accused the company of “banning free speech” in coordination with “the Democrats and Radical Left”,’ the new study reads.

The study, which has not been peer-reviewed, shows evidence that conservative politicians and media receive as much attention as their liberal counterparts — if not more.

Contrary to this idea, Trump has been by far the center of social media attention in the 2020 elections. Even as he demonized big tech in 2016, this very big tech (and social media especially) were largely responsible for his victory. Brad Parscale, his former digital campaign chief, told Wired just after the election:

“Facebook and Twitter were the reason we won this thing. Twitter for Mr. Trump, and Facebook for fundraising,” adding that Targeted ads on Facebook helped generate the bulk of the $250 million the first Trump campaign raised online.

Trump dominated online discourse, both in the 2016 and the 2020 elections, write Paul M. Barrett and J. Grant Sims, the authors — and it’s not just Trump. A previous 2020 study found that conservatives in general ruled online discourse for this hot topic, with users sharing the most viral right-wing content about Black Lives Matter about 10 times more than the most popular liberal posts on the topic. A good example is the Black Lives Matter Facebook page, which has 0.7 million fans — far less than the Blue Lives Matter page set in response to it. “Even anecdotal evidence of supposed bias tends to crumble under close examination,” the researchers write.

The study looked at millions of posts from Facebook, Instagram, Twitter, Reddit and 4chan, but that hasn’t stopped leading Republican members from blaming social networks of bias.

“What we are watching — the militarization of social media on behalf of Democrats, and the overt suppression of material damaging to Democrats to the cheering of the press — is one of the single most dangerous political moments I have ever seen,” conservative commentator Ben Shapiro wrote on Twitter on Oct. 15. Shapiro has 3.3 million followers on Twitter.

This alleged bias does not seem to be backed by facts, the authors write. Instead, as Renée DiResta of the Stanford Internet Observatory describes it, this bias allegation is nothing more than a ‘narrative, deployed as a cudgel by politicians who use it cynically to rally their base’ — and it works.

A survey from Pew Research Center in August indicated that 90% of Republicans believed social media sites censored political viewpoints. Even 59% of Democrats had the same view. Things aren’t always straightforward, but Americans seem to believe social media acts as a censor — and the facts don’t back the narrative.

In fact, social media doesn’t seem to be targeting conservatives or Republicans, but rather “people who violate its rules by calling for violence, harassing others, or advocating hateful ideologies,” the study reads. Among the right-leaning users who have faced enforcement action on Twitter are white nationalists like Richard Spencer, Jared Taylor, and David Duke, as well as white nationalist organizations such as the American Nazi Party, the neo-Nazi Traditionalist Worker Party, and American Renaissance magazine. In other words there’s no problem if you’re a conservative, but if you’re leaning on a neo-Nazi party or white supremacy, or you’re advocating for violence, you may get a Twitter ban.

This is not to say that social media platforms are perfect — far from it. Big Tech is flawed in many ways, and this bias allegation comes in large part due to the lack of transparency of social media networks. There is certainly a lot of work that needs to be done on multiple fronts, and calls for increased transparency and regulation are only getting stronger.

In fact, the European Union has announced plans to regulate big tech companies more tightly, but since one of the main focus points is removal of ‘harmful content’, that may end up with more bans for hateful content, not less.

Ultimately, while it’s hard to both prove and disprove any type of social media bias, there doesn’t seem to be much evidence (if any) to support this alleged bias. If anything, it seems more like the bias allegation itself is disinformation, the researchers conclude.

“The claim of anti-conservative animus [on the part of social media companies] is itself a form of disinformation: a falsehood with no reliable evidence to support it,” the NYU researchers write.

“[..]as they condemn supposed social media favoritism, conservatives appear to relish wielding
the bias-claim cudgel, even though it’s based on distortions and falsehoods.”

A long-lasting problem

Trump has been the leading purveyor of this accusation, and his vitriolic rhetoric is unprecedented for an American president. But the narrative won’t disappear with him. Almost certainly, right-wing channels (and far-right-wing channels) will continue to push it, especially because it resonates so well with some voters. So what should be done?

The researchers propose a list of 7 actions that social media companies can take to increase transparency and become more trustworthy:

1. Provide greater disclosure for content moderation actions. Platforms don’t typically provide much justification for why a post or account is sanctioned, and sometimes, the results are perplexing. For instance, conservatives protested when Twitter flagged Trump for promoting violence, but didn’t do the same for Ayatollah Ali Khamenei, Iran’s supreme leader, who threatened Israel with annihilation. This is not a sign of anti-conservative bias, but it is a lack of transparency and consistency.
2. Offer users a choice among content moderation algorithms. Some users want to see more violent and hateful content. Others don’t. Giving users the chance to choose between these two moderation variants would empower them and also reveal how some of the darker social media bubbles form.
3. Undertake more vigorous, targeted human moderation of influential accounts. Big tech companies like to say they leave moderation to the algorithms, but human moderation is severely missing from these platforms. At least for high-profile cases, human moderation needs to be more vigorous. Of course, for this, tech companies would have to hire and train more employees.
4. Release more data for researchers. Part of the reason why it’s so hard to disprove allegations of an anti-conservative myth is because the available data isn’t sufficiently detailed. Social media platforms offer some data, but it’s not enough to carry out detailed analysis, researchers point out.
5. Pursue a constructive reform agenda for social media. Big tech companies are here to stay, at least for the foreseeable future. Trying to destroy them won’t do anyone any good — instead, the White House should focus on a reform and not wait years for court battles. “The president—perhaps by empowering a special commission—needs to foster engagement with the private sector on content related issues,” the researchers write.
6. Work with Congress to update Section 230. Enacted in 1996, this law helped online platforms and protected them from a swath of potential liability — which was useful when they were small companies with small budgets. Now, as these companies have grown and changed so dramatically, their liability needs to change accordingly.
7. Create a new Digital Regulatory Agency. Some of the distrust in tech companies is warranted. But another part of it comes directly from the proliferation of myths such as the anti-conservative bias. Trust is extremely hard to build, but researchers suggest that a new independent body that would enforce the updated changes in Section 230 could help wtih that.