In the final weeks ahead of the 2018 midterm elections, as Republican seats looked increasingly imperiled, many campaigns and activists reported a flood of online attacks on progressive candidates – often misleading, savage, and incendiary. But confusion reigned: Which ones were real and which were from trolls and bots? Who was really behind these online attacks? Campaigns needed to understand their online threat landscape quickly to allocate resources and prevent voter suppression and deception.
GQR rapidly processed over 500,000 social media posts relevant to races and voters in Georgia, Florida, Michigan, and Wisconsin. Our analysis identified key threats including artificial activity and coordinated online attacks, as well as the level of threat, based on a proprietary social media threat index specific to each race and voter landscape. Our analysis helped independent expenditure efforts prioritize resources around voter outreach and digital organizing in the final weeks ahead of Election Day, and demystified meaningless online activity that threatened to distract from more genuine online threats.
With access to social media data through a dashboard fire hose and historical knowledge to expedite manual listening and pattern recognition among publicly available social media data on platforms with more restricted access like Facebook and Instagram, GQR conducted network and threat analysis to identify coordinated digital activity. We successfully informed voters about false, misleading, and divisive content, and pulled back the curtain to expose influential artificial accounts attempting to impact elections.