Russia 2016 election interference: What’s real, what’s overhyped?

Russian trolls were overhyped.

This is the implication of a new study in Natural Communication, written by a team of six academics who tried to assess whether the Russian government’s Twitter propaganda efforts during the 2016 campaign actually changed the minds of users. “We found no evidence of a meaningful relationship between exposure to Russian foreign influence campaigns and changes in attitudes, polarization, or voting behavior,” the authors wrote.

This is not a surprise to me – I have long believe Russian troll farms have little impact. But the aftermath of the Trump-Russia scandal remains fiercely contested — with many on the right and some on the “heterodox” left continuing to question it. did Russia do anything important — it’s worth looking back and taking stock of what the Russian government did that year. Because that’s okay.

Basically, the Russian government tried to interfere in the 2016 election, and it did so in two main ways.

First, there is the social media propaganda effort: Russian trolls. On Facebook, Twitter, and Instagram, Russians impersonate Americans and post content designed to inflame US political tensions, often attacking Hillary Clinton and supporting Donald Trump. This is the case, and it is somewhat unusual, but as this new study confirms, there is no reason to believe that it makes any difference to the voting behavior of Americans. Russian propaganda is just a drop in the sea of ​​media and social media messaging launched by America.

Second, and even more so, there is the stolen email: hack-and-leak. Russian intelligence officers hacked and obtained emails and documents from many top Democrats and have publicly disclosed them — giving some to WikiLeaks, giving others to journalists, and posting more on websites they control.

The hack and the release made an impact — in particular, it was a negative story for Clinton that continued throughout the last months of the campaign. But did they swing the election? That, in a counterfactual world without Russian interference, Clinton would have won? My assessment is “probably not,” but it is difficult to conclusively say for sure.

Still, this intervention is quite unusual in the context of American elections; it is appropriate to treat them as a big problem and it is understandable that Americans can be angry with Russia trying to sway their votes. There are real victims here, in people who see personal correspondence thrown on the internet, and it is always important to prevent this kind of intervention from happening in the future.

Russian trolls did not swing the 2016 election

Russia’s social media propaganda efforts were conducted by the Internet Research Agency, an organization based in St. Petersburg. war). The IRA is not officially a government agency, but a bipartisan report from the Senate Intelligence Committee concluded that “the Russian government engaged in and supported” the 2016 election meddling effort.

IRA employees create online accounts, claim to be Americans, and post inflammatory material or trolls about US politics. Much of this material is rather crude or stupid, but the overall thrust is clear: It tends to praise Trump and attack Hillary Clinton. The IRA also paid for online ads featuring the following pro-Trump, anti-Clinton message, as seen in this table from the indictment brought by Robert Mueller’s team, which is tasked with investigating Russian interference with the election:

Indictment of the Internet Research Agency, US Department of Justice

All of this was a source of media attraction and attention when the news broke after the election. Facebook estimates that IRA messaging reaches up to 126 million people on its platform, which certainly seems like a huge number. “It makes all of us who use social media to keep up with friends, share photos and follow the news thinking: How did the Russians find me?” Geoffrey Fowler of the Washington Post wrote in November 2017.

But there are certainly strong reasons to doubt that IRA remittances have had much of an impact, even if the numbers are scary. If I scroll through Twitter for just 15 minutes, I can be “reached” by messages from hundreds of sources, if not more. I’ll use the metaphor again that this is a drop in the ocean: Americans have been swimming in messaging about the 2016 election from all kinds of traditional media and social media sources for months (as well as their campaigns and ads), and there is no reason to believe any of the posts or ads. Russia has a special persuasive power that all other messages lack.

Now, the author of Natural Communication study – Gregory Eady, Tom Paskhalis, Jan Zilinsky, Richard Bonneau, Jonathan Nagler, and Joshua A. Tucker – have conducted their own analysis driving this front. Back in 2016, they surveyed nearly 1,500 US Twitter users at various points in the campaign, returning to the same respondents to track how their opinions changed. They also analyzed all the Twitter accounts followed by the respondents to see how many were exposed to Internet Research Agency content.

Overall, they found that 70 percent of exposure to IRA posts were concentrated among 1 percent of respondents, and those respondents are usually highly partisan Republicans who have favored Trump. Which makes sense, knowing about social media – if you’re looking for anti-Clinton, pro-Trump content, you’ll find it, because the algorithm’s recommendations or the accounts you follow just retweet the agreed message without really knowing who said it. And that’s the kind of content IRA provides. But found suggesting IRA messaging is not well calibrated to target wavering swing voters.

The authors also put this messaging exposure into perspective by finding that users had more posts from traditional media and politicians on their timelines than Russian trolls — “an order of magnitude” more, they wrote. And finally, the results of the survey “did not detect a meaningful relationship between posts from Russian foreign influence accounts and changes in respondents’ attitudes on issues, political polarization, or voting behavior.”

Now, it is true that this is only a study of Twitter posts and not other social media platforms like Facebook where the IRA is also active. The overall logic is sound, though. Americans can certainly be outraged by Russia’s efforts to spread its views, and Mueller’s team says much of what it does violates U.S. law. But Russia doesn’t have some magical messaging that sways the minds of Americans and compels them to support Donald Trump.

Hack-and-leak is a bigger problem (but still might not swing an election)

In 2016, the GRU, Russia’s foreign military intelligence agency, hacked Americans involved in politics or government.

Using a simple spear-phishing email (a link to a fake Google page asking people to enter a password), they gained access to the email accounts of Clinton campaign chairman John Podesta, as well as several Clinton staffers, volunteers, and advisers. According to Mueller’s report, they spear-phished an employee of the Democratic Congressional Campaign Committee, who obtained credentials to access the DCCC’s computer network and eventually connected the Democratic National Committee’s network. Hackers then use malware to retrieve emails and documents.

(Some critics of the Russia investigation like journalist Matt Taibbi still considered skeptical that the hacks were the Russian government, but the Mueller report and the indictment of 12 Russian intelligence officers provide a wealth of specific detailed allegations about exactly how these hacks happened and which specific officers and units in the GRU are responsible. This is inconvenient for Russiagate skeptics, so they tend to ignore it.)

Foreign hacking is far from unusual; The Chinese government, for example, is said to have hacked the presidential campaign of Barack Obama and John McCain in 2008, and the US is likely to have done the same amount as well. The shocking denial is what happened next: The stolen information began to appear publicly, in large accounts. (Another unusual aspect is that one of the main candidates openly welcomes this intervention, with Trump invite the public Russia to “find” more Clinton emails.)

The GRU used the persona “Guccifer 2.0” (“Guccifer” was the name used by the imprisoned Romanian hacker) and registered a website, DC Leaks, posting the hacked material there and providing some of it to journalists. But two groups are specifically reserved for WikiLeaks, a non-profit organization that has posted leaked US government material in the past.

First, in late July 2016, before the Democratic National Convention, WikiLeaks released thousands of DNC emails and revealed that many DNC members privately spoke of Bernie Sanders with disdain. The revelation prompted the resignation of DNC Chair Debbie Wasserman Schultz and other top staffers, and the overall start of a bad Democratic convention.

Second, in early October 2016, WikiLeaks began sending Podesta’s emails – and will continue to send them, in batches, until the election. Podesta’s emails aren’t as explosive, but in anyone’s private email there will be embarrassing material that’s preferable, and if that embarrassing material is released only about one candidate, it could have an impact. So each new release generated some media coverage, and they were mentioned continuously by Trump on the campaign trail, who often baselessly claimed that they revealed some malfeasance or another.

Trump sank in the polls for the first half of October, as the controversy over him Access Hollywood tape hit the news. In the second half of the month, the news cycle shifted and his poll numbers rebounded. He didn’t overtake Clinton in the polls, but he came close — close enough to pull out a victory by less than 1 percentage point in three key swing states.

But I would be hesitant to claim that Podesta’s email coverage drip is responsible for Trump’s recovery. Part of it can only be about the Republican partisans temporarily killed by the scandals that “come forward” before the election – a process that may have unfolded even if there had been no Russian interference. (Trump also did well in the polls in the latter half of October 2020.)

The case can be made, just by keeping the word “Clinton” and “email” in the news for a month, Podesta’s release damaged her campaign (because the leaked coverage will remind voters of a separate matter, the FBI investigation into Clinton’s email). Still, it’s far easier to make the case that FBI director James Comey last October’s letter said the new Clinton email has been found changed successfully, as Nate Silver has argued, from the Russian hack-and-leak. Comey’s letter was a discrete event that dominated the headlines and preceded a sharp shift in the polls, while the Podesta leak was one of the stories in the background throughout October.

So I was never convinced that the Russians had fully elected Trump. That’s a high bar, and despite the overhyped Russian trolls, it seems like the hack-and-leak was an important intervention that hurt Clinton and the Democrats to some extent. And indeed there is an effort to influence the American vote and undermine a presidential candidate that Russia doesn’t like – an effort that violates US laws against hacking and disclosing campaign activities. The intervention was real, the investigation was justified, and the Russian government was outraged at what it was trying to do.