The pandemic has shut almost everything, and people are spending more time online. Compared with 2016, the 2020 presidential campaigns will probably engage more with voters on phones and laptops, rather than in-person rallies. So does this mean we’re likely to see greater foreign interference in this election?
There are signs that this type of interference already was happening. In February, U.S. intelligence officials briefed Democratic presidential candidate Bernie Sanders that Russia was trying to boost his campaign — echoing a parallel effort in the 2016 election.
Another example surfaced last fall, when Facebook took down the “I Love America” page, which had 1.1 million followers and within three months had attracted more online interactions than most U.S. media companies combined. Although there is no indication that any government backed this page — the content managers were Ukrainians — this Facebook page recycled material produced before the 2016 election by a Russian troll farm, the Internet Research Agency (IRA).
The TMC newsletter is changing shape! Sign up here to keep getting our smart analysis
What do Russian trolls want, and what’s their strategy?
In a recent report, we examined whether the IRA-led accounts used different strategies in 2016 when targeting left- vs. right-leaning U.S. voters. Our main takeaway is that the Russian campaign attempted to agitate conservative audiences while building loyal followings among liberals.
Here’s how we did our research. We selected a random sample of 898 memes across 45 popular IRA-curated Instagram, Facebook and Twitter accounts from the @UsHadrons Medium repository (a collaboration of citizen journalists and Josh Russell). That’s about 10 percent of the repository’s archived memes.
This covid-19 misinformation went viral. Here’s what we learned.
The repository tracks 50 accounts generating content that circulated during the 2016 U.S. election. The creators of the repository did not provide their selection criteria, so we verified all but four accounts against the House IRA ad archive (30 accounts verified), the House list of IRA Twitter handles (seven), the Senate Intelligence Committee report (four), news accounts (two) or the New Knowledge report (two). We also excluded one verified Russian account — Don’t Shoot — because it promoted videos but no memes. Our report provides further details on the methodology.
These 45 accounts don’t comprise the universe of IRA accounts: the House Committee on Intelligence identifies 470 IRA Facebook pages and 3,841 Twitter handles (including 123 primary and 3,718 amplifier accounts). The New Knowledge report identifies 133 IRA Instagram accounts.
Our subset of 45 Russian-led “users” merits attention because these accounts generated the vast majority of viral content in the Russian disinformation campaign. This content includes the top 20 IRA Facebook pages that generated 99 percent of all IRA audience engagements. It also includes 17 of the top 25 IRA Instagram pages that accumulated approximately 64 percent of all IRA engagement on Instagram. And it includes six of the top 20 IRA Twitter handles that generated 65 percent of retweets by the 20 top Russian Twitter accounts
We coded each account for its general positive or negative tone, looking for statements of pride or grievances — and logged the dominant theme (e.g., pro-LGBT or pro-gun). We looked at the intended audience, gauging whether the account targeted left- or right-leaning voters. We also coded each of 898 memes according to the types of images pictured, and classified each meme by the type of ideological message it portrayed.
Here’s what we found
Overall, there were 20 left-leaning accounts (18 pro-minority groups, one anti-conservative, and one anti-police) and 25 right-leaning accounts (15 anti-liberal, three pro-veteran, two pro-police, two anti-immigrant, one pro-gun, one anti-all government, and one anti-Hillary Clinton).
We determined that 75 percent of the right-leaning accounts had a negative tone, while 90 percent of left-leaning accounts had a positive tone. Among the right-leaning content, 68 percent included at least one of the recorded images (a U.S. flag, religious object, weapon, politician or celebrity). By contrast, 75 percent of left-leaning memes contained none of these items. Of note, 23 percent of right-leaning memes displayed politicians, compared to only 2 percent of left-leaning memes.
Other studies find similar trends, including a divergence in the substantive messaging of right- and left-side trolls, and the IRA infiltrated both sides of the #BlackLivesMatter online discussion, for instance.
Here’s what this tells us about Russian recruitment methods
Our analysis offers some new insights — we found that while the IRA used negative, politicized content to target conservatives, it also attempted to create loyal followings among left-leaning voters through a three-step formula: affirming the (minority) identity; avoiding electoral politics, and alluding to well-established group grievances.
Americans are fighting coronavirus misinformation on social media
This tactic of trying to build identity-affirming online communities for minority groups is a continuation of past Soviet attempts to exploit U.S. racial tensions. The 2019 Mueller Report describes the 2016 efforts as “Targeting and Recruitment of U.S. Persons” (pp. 31—33) and “U.S. Operations Involving Political Rallies”(pp. 29—31). These efforts aim to draw in Americans online, effectively exploiting the regular political process without subverting it.
A Wall Street Journal investigation last year revealed that eight IRA Facebook pages scheduled 60 U.S. rallies, at least 22 of which took place. Of those 60 rallies, 19 supported conservative causes and three supported liberal causes. The article mentions “sparsely attended” conservative gatherings and liberal events with more participants — including 300 attendees protesting the police shooting of Philando Castile near Minneapolis, and dozens at an Orlando LGBT vigil after the Pulse nightclub shooting.
These accounts, along with our research, suggest the IRA probably scheduled a greater number of conservative rallies, yet their liberal events mobilized more individuals. This underscores how loyal and engaged the followers of liberal-leaning IRA accounts were in 2016.
In-person protests have become smaller in many U.S. states during the pandemic lockdown, but Russia’s online attempts to engage voters no doubt will continue in the lead-up to the November 2020 presidential election. These efforts look to affirm group identities, engage with voters on apolitical topics or reinforce well-established group grievances. Whether the IRA stumbled upon the differing patterns of targeting voters or designed them, these approaches show Americans of any ideology may become “willing but unwitting collaborators” in future foreign influence campaigns.
Brigitte Hugh is a master’s student in political science at Utah State University.
Anna O. Pechenkina is an assistant professor of political science at Utah State University.