Amid the generalized media crack-up that surrounded the 2016 presidential election, the bogeyman of “Russian bots” quickly became a load-bearing concept. A Russia-based social media campaign, or so it was said, had saturated sites like Twitter with fake accounts and, in doing so, helped to swing the election for Donald Trump. Becoming axiomatic in liberal circles, this story soon took on a life of its own. It’s since played a prominent role in mainstream media narratives of the 2016 election, been the subject of highly publicized congressional hearings, and also loomed large in the wider global discourse about “fake news.”
That the Russian government preferred Trump to Hillary Clinton and that Russia-connected actors engaged in digital skulduggery related to the election are not really in dispute. Much of the mainstream discussion around Russian bots, however, has been premised on unexamined assumptions about the scale and effectiveness of these efforts. Powerful states including the United States, after all, regularly engage in the likes of online propaganda and sock-puppeting campaigns. Whether they have a more than negligible impact on real world events, electoral and otherwise, is another question.
It’s notable, then, that a new analysis published by the Center for Social Media and Politics at New York University finds no evidence whatsoever that Russia-based Twitter disinformation had any meaningful impact on voter behavior in 2016. In place of the terrifying bot army menace that’s periodically been invoked, the researchers instead detail an enterprise with minimal reach or influence, and one overwhelmingly concentrated among partisan Republicans already inclined to vote for Trump.
They estimate that as many as thirty-two million US Twitter users may have been “exposed” to tweets from Russia-aligned accounts over the eight-month period preceding the 2016 election.” In numerical terms that may sound like a lot, but it actually isn’t when you factor in the sheer volume of posts and information encountered by social media users on a daily basis. As the report puts it:
While, on average, respondents were exposed to roughly 4 posts from Russian foreign influence accounts per day in the last month of the election campaign, they were exposed to an average of 106 posts on average per day from national news media and 35 posts per day from US politicians. In other words, respondents were exposed to 25 times more posts from national news media and 9 times as many posts from politicians than those from Russian foreign influence accounts.
Sheer exposure, of course, doesn’t even necessarily amount to influence. Like advertising, politically motivated content can functionally be background noise if it fails to reach particular audiences or in turn doesn’t have an impact on those that it does. In both respects, the study is quite unequivocal: not only were Russian Twitter efforts dwarfed by posts from media and politicians, but actual exposure to them was highly concentrated within a subset of partisan conservatives:
Results . . . show that the amount of exposure depends substantially on users’ self-identified partisanship: those who identify as “Strong Republicans” were exposed to roughly nine times as many posts from Russian foreign influence accounts than were those who identify as Democrats or Independents.
Even then, in fact, it was unlikely to produce changes in attitude or behavior. As the researchers conclude:
We 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.
While there are, as the study’s authors hasten to note, some limits to the analysis (it’s confined to Twitter posts as opposed to other kinds of media content or posts on other networks) it nonetheless offers persuasive evidence that the “Russian bots” narrative of the 2016 election has greatly overstated the impact of Russia-based social media efforts on the outcome. The result has been an account of 2016 sometimes quite unmoored from reality, wherein what was in practice a relatively impotent operation has elicited a torrent of angst-ridden and often frenzied media coverage.
If the Russian bots story gained momentum without having much of an empirical foundation, one reason is that it offered traumatized liberals a tidy and straightforward explanation for an outcome they had spent the preceding year believing was impossible. Sinister as the idea might be, a foreign campaign of digital sorcery was always going to be a neater culprit than the litany of institutional and political failures that actually enabled Donald Trump to become president.
In a different kind of world, November 2016 might have inspired some actual introspection on the part of those implicated in said failures. Instead, it became an occasion for hyperbolic and often flimsy partisan narratives that rather conveniently avoided asking the necessary questions — and in turn let the triangulating ideology of Clintonite liberalism off the hook.