A new study conducted by New York University has concluded that Russian influence operations on Twitter played no measurable role in Donald Trump’s victory in the 2016 election. It’s a striking finding, and one that should prompt some reflection. Some Democrats grew obsessed with Russian meddling in American affairs during the Trump era, to the point where they questioned the election’s legitimacy.
But while Russia’s goal of interfering with the U.S. elections has long been clear, what’s been less clear is whether it actually worked. The NYU study demonstrates that Russia’s influence operation on Twitter was a dud. Those who feared a new age of foreign countries shaping U.S. elections would be wise to consider the difference between Russia’s intentions and its actual effectiveness, and to contemplate how overestimating Moscow can distort their understanding of what’s happening at home.
According to the study’s authors, an estimated 32 million Twitter users in the U.S. “were potentially exposed to posts from Russian-sponsored accounts in the eight months leading up to the 2016 election.” Researchers and the U.S. intelligence community believe Moscow used outfits like the Internet Research Agency — basically a professional internet troll factory — to try to encourage Americans to either vote for Trump or to discourage them from voting for the Democrats, in some cases by encouraging disillusioned Bernie Sanders supporters to either not vote or back a third party. Analysts also believe that the Russians were looking to generally increase distrust and polarization in the American political system.
The study, published in the journal Nature Communications, uses longitudinal survey data matched with Twitter feeds to assess exposure to Russia’s influence campaign and whether it affected their attitudes or voting behavior. What the study found was that Russia's campaign wasn’t effective for a number of reasons. First, only 1% of Twitter users accounted for 70% of all exposures to foreign influence accounts, and the highest numbers of exposures were concentrated among partisan Republicans. Second, the study found that influence operation tweets were dwarfed by content from domestic media and political sources. And, most damningly, the authors of the study “did not detect any meaningful relationships between exposure to posts from Russian foreign influence accounts and changes in respondents’ attitudes on the issues, political polarization, or voting behavior.”
This study does not rule out that Russia’s interference influenced the 2016 election. This study focused only on Twitter, and it’s possible that Russian operations were more effective on Facebook, where there were more users and more exposures. (That said, when asked to guess how Russia’s influence operations on Facebook might’ve compared to Twitter, study co-author Joshua Tucker told me that he suspects that troll operations on Facebook could be similarly ineffective to those seen on Twitter, given broader trends in influence operations.) And according to intelligence analysts, Russia also attempted to influence the U.S. election by hacking into the computers of the Democratic National Committee staffers and the Clinton campaign. The resulting leaks were then covered extensively by the U.S. press.
But, overall, this study deals a big blow to the narrative that Russia may have thrown the election for Trump. It should also temper Democratic lawmakers’ post-2016 anxieties that Russian social media operations have demonstrated themselves as a grave threat to American political life.
One of the implications of this study is that American journalists might bear a greater share of the blame in Russia’s influence operation than we previously thought. Here’s why: If Russia’s social media influence operations are far less effective than commonly believed, then whatever effectiveness Moscow’s meddling did have was related to the hacks and leaks of the DNC and the Clinton campaign. As political commentator Matt Yglesias has rightly pointed out, it must be remembered the participation of the American press was essential for those operations to work. Journalists of course have every right to use leaked information that’s been released to the public. But as Yglesias argues, the centrist press’s collective decision to frame the leaks — which included unflattering information about Hillary Clinton — as symmetrical to information about Trump’s corruption was a serious misjudgment. There is a lack of rigorous data to determine the fallout of that misjudgment, but it’s worth closer examination.
This study also serves as a reminder of the perils of conflating Russia’s adversarial intentions with its actual abilities. If Moscow’s influence operations are understood to be a lot clunkier than we thought, then it puts a greater burden on Democrats to figure out what was going on in the domestic political scene that made MAGA politics appealing, and forces them to reckon with the reality that U.S. citizens in the voting booth were not just foreign-influenced dupes.
Lastly, the study should hopefully help reduce concerns about the legitimacy of the 2016 election — and the electoral process in general. It’s important to clearly understand what does and doesn’t seriously threaten American democracy. In a time of declining trust in the democratic process, it’s crucial to not unnecessarily inflate threats about how easily the system can be manipulated.
None of this is meant to downplay the impropriety and national security risks posed by Russian interference in the election. Nor does it rule out the possibility that Russia — or other foreign adversaries — could act with far greater sophistication in the future. But Russia isn’t all-powerful, and we already know for certain that Americans are doing a fine job as it is feeding each other home-grown manipulation and misinformation.