Relevant Papers:

This study analyzed tweets from handles associated with the Russian Internet Research Agency in an effort to better understand the tactics employed by that organization on the social media platform Twitter in their attempt to influence U.S. political discourse and the outcome of the 2016 U.S. Presidential election. We sampled tweets from the month preceding the election and analyzed to understand the qualitative nature of these tweets as well as quantitative differences between how types of IRA Twitter accounts communicated. Seven categories of tweet behavior were identified: attack left, support right, attack right, support left, attack media, attack civil institutions, and camouflage. While camouflage was the most common type of tweet (52.6%), descriptive analyses showed it was followed by attack left (12%) and support right (7%). A variety of quantitative differences were shown between how account types behaved

From 2015 through 2017, the Russian-government affiliated Internet Research Agency produced nearly 2.8 million English-language tweets from accounts that purported to be operated by    U.S. nationals or organizations (“trolls”). Almost half of the trolls’ output were retweets of or replies to other accounts, overwhelmingly from outside the network. We analyze the characteristics of outside accounts that were targeted by the trolls in this way, and how this behavior changed over the life of the operation, in order to infer what role contacts with outsiders played in the trolls’ propaganda strategy. We document the three stage life-cycle of these externally-oriented trolls: introduction, growth, and amplification. In a quasi-experiment of the amplification stage, we estimate that in the month leading up to the 2016 U.S. Presidential election, the trolls induced about 3 million additional tweets from and 4 million additional followers for the 25,000 unique accounts they amplified, an impact that rivals the direct output of the trolls themselves over the entire three-year campaign.

We document methods employed by Russia’s Internet Research Agency to influence the political agenda of the United States from September 9, 2009 to June 21, 2018. We qualitatively and quantitatively analyze Twitter accounts with known IRA affiliation to better understand the form and function of Russian efforts. We identified five handle categories: Right Troll, Left Troll, News Feed, Hashtag Hamer, and Fearmonger. Within each type, accounts were used consistently, but the behavior across types was different, both in terms of “normal” daily behavior and in how they responded to external events. In this sense, the Internet Research Agency’s agenda-building effort was “industrial”– mass produced from a system of interchangeable parts, where each class of part fulfilled a specialized function.