Biased NYU 'research' calls claims of censorship against conservatives 'disinformation'

Researchers from New York University's Stern School of Business released a study on Monday arguing that claims of conservatives being censored more than liberals on social media amounts to "disinformation."

The study itself, however, is riddled with bias against conservatives and bares a flawed methodology which raises questions about the accuracy of the study.

A biased introduction and breakdown

Already by the second paragraph, the authors make a biased and unsubstantiated generalization about conservatives being conspiracy theorists. "This accusation—that social media platforms suppress conservatives— riles a Republican base that... is prone to seeing public events as being shaped by murky liberal plots," the study reads.

"But the claim of anti-conservative animus is itself a form of disinformation: a falsehood with no reliable evidence to support it," the study continues in the next paragraph, further arguing that examples of conservatives being banned from social media for unjustified reasons amounts to "anecdotal evidence."

A number of problems already appear from such a statement. First of all, while the crux of evidence that conservatives are being censored on social media may be anecdotal in absence of reliable studies on the matter, a lack of evidence for a proposition does not amount to evidence against the proposition.

Second, even a single study on the matter is not enough, in scientific analysis, to declare a proposition a "falsehood," and certainly not enough to call it "disinformation" when anecdotes can be used to back up a claim. Disinformation is defined as "false information which is intended to mislead," and while one can argue that the claims that conservatives are being censored is overblown or untrue, it would be very difficult to argue that it constitutes an intention to mislead given the various anecdotes of conservatives who have been censored by social media platforms without breaking their rules.

Furthermore, the study goes on to excuse the banning of certain high-profile right-wingers from various social media platforms, including former President Donald Trump, which they described as a "reasonable attempts to forestall additional violence." They justified this on the basis that some, more violent Trump supporters had interpreted recent tweets of his as supportive of violence against the government, even though he never made any explicit statement calling for violence, as is the standard set by the first amendment.

While one might argue that social media companies are not bound to restrict speech solely on the basis of the first amendment, that is not the argument being made. Whether social media sites are justified in censoring someone outside the bounds of the first amendment is an entirely different question from whether their actions constitute censorship at all. Ultimately, the argument that social media sites are justified in censoring opinions they dislike is irrelevant to the question of whether they censor or not. While the authors of the study may view the social media censorship as justified, it ultimately detracts from their argument that social media companies do not engage in censorship of conservatives.

The study also justifies the ban of Alex Jones from social media platforms due to his spreading of disinformation regarding the Sandy Hook school shooting. It is worth noting this in the context of the study describing claims of anti-conservative censorship itself being "disinformation." Under the framework of this study, censorship of those who claim that social media companies censor political opposition would be considered justifiable on the grounds of spreading "disinformation."

To make this clearer, under the framework of this study, banning people from social media for alleging censorship by big tech would not be considered censorship by big tech. If the methodology of this study does not sound questionable already, wait until you hear what the actual methodology is.

The study also justifies a number of other notorious examples of social media censorship, most notably of the New York Post's Hunter Biden bombshell which was released in October of last year. The study describes the censorship of the story as a "reasonable decision" as the story itself was "questionable."

The authors of the study did not even bother to dispute the actual claims of the study on a factual basis, and these mere dismissals served as their excuses for justifying the censorship of this story.

It should also be noted that the censorship of the Hunter Biden story cannot simply be dismissed as anecdotal. While it may be a single story, one survey found that one in six Biden voters would have switched their votes had they heard about the Hunter Biden allegations, far more than enough to swing an election. Even if the survey, which was conducted by a conservative-leaning institution, was itself biased, Biden narrowly won the election by a margin of less than 43,000 votes in three states. To suggest that the censorship of the story is merely an "anecdote" as opposed to a a highly scandalous affair which may have swung an election is ludicrous.

Poor methodology

The methodology of this study does not say anything whatsoever about whether conservatives are censored by tech giants.

"Data from a variety of sources provide a window into what Facebook posts get the most engagement, meaning likes, shares, comments, and other reactions," the study says. "If engagement is a rough proxy for popularity, then the right wing is holding its own."

Earlier in the paper, the researchers made the argument that conservatives being more likely to be banned on social media does not in and of itself indicate censorship. This is a true claim, as correlation does not equal causation, and it could simply be the case that conservatives are more likely to violate the rules of social media platforms. The only conclusion which can be definitively confirmed from conservatives being banned more often on social media is that conservatives are banned more often.

The researchers do not appear to realize how such disparities between correlation and causation apply to their own study, however.

"On most days, right-leaning U.S. Facebook pages dominate the list of sources producing the most-engaged with posts containing links," the study reports, noting that pages such as Breitbart, Bongino Report, Fox News, and Daily Caller all consistently outperformed liberal and left-wing news outlets.

However, the fact that right-wingers outperform left-wingers in social media engagement does not, in and of itself, disprove claims of censorship. It could simply be that conservatives are better at using social media, or that they are more likely to use it. However, the only thing which can be definitively confirmed by right-wingers outperforming left-wingers in terms of social media engagements is that right-winger outperform left-wingers on social media platforms. No other conclusion can be definitively identified by this fact.

I wrote a single empirical research paper while studying in university as an undergraduate economics student, but even I could think of a better methodology to research claims of anti-conservative social media censorship. You would start by taking a random sample of various politically-oriented social media users, perhaps including a small follower floor to control for fake and/or troll accounts. Next, you would categorize the posts, comments, replies, etc. of their accounts by political affiliation, whether they appear to have broken platform rules, and whether they had moderative action taken against their accounts over the course of a month or two.  You would then split the results by other potentially relevant factors such as gender, number of followers, etc.

It's not the most sophisticated study, but at least it would go beyond correlation to determine the censorial bias of social media companies. Apparently, however, the authors of this study could not, and believed they could make wide-ranging conclusions solely from these questionable findings.

The media response

Of course, the embarrassingly incurious demographic which makes up the bulk of mainstream media journalists parroted the results of the study without question.

"Republicans including Donald Trump have raged against Twitter and Facebook in recent months, alleging anti-conservative bias, censorship and a silencing of free speech," reported The Guardian. [None] of that is true."

"Baseless claims of anti-conservative bias are driving Republicans' approach to regulating tech," reads the Washington Post. Meanwhile, Forbes said that the study "debunks a series of right-wing claims as false or based on unsubstantiated evidence."

Not a single one of these journalists seemed to have noticed the incredibly flawed methodology of these studies, assuming they even read them at all. Most of them actually do talk a bit about the methodology though, without any mention of the very questionable conclusions the researchers are trying to draw on it. Certainly none of them identified the very classic and common correlation and causation problem. As I said, embarrassingly incurious.

It should not, however, be surprising to see such positions from the mainstream media. Left-wing and liberal media outlets have consistently supported greater censorship on social media, often leading the call. So it's no surprise that they would uncritically report a study which both justifies censorship while gaslighting you into thinking it's not happening at all. CNN's Brian Stelter did it in the span of two minutes just one day before this study was published.

Ultimately, the censorship will likely continue until the range of acceptable views gets narrower and narrower. At that point, such gaslighting from journalists and academics whose views are too orthodox to face censorship will not be enough to quell the concerns of the politically diverse, censored Americans.