The next speaker in our AoIR 2022 session on elections is Fabio Giglietto, and focusses on political advertising and coordinated behaviour in the lead-up to the 2021 German election. Sponsored by the Media Agency of North-Rhine-Westphalia, it was interested in micro-targetting of ads on social media as well as coordinated behaviour, and proceeded by identifying the social media accounts of a large number of candidates in the German election. It also worked with a list of relevant political terms compiled by GESIS.
This enabled the project to gather relevant content from Facebook, Facebook ads, Twitter, Instagram, and the researchers then used the CooRNet tool to explore coordinated link-sharing behaviours on social media during the election. By identifying URLs in social media posts and then exploring whether these URLs were shared repeatedly within very brief timeframes, it is possible to detect suspicious coordinated activity and explore the networks of accounts that seem to be driving this.
Over the six weeks of the election campaign, this detected a number of networks of coordinated link-sharing. It identified networks of organised posting by politicians in the same party; editorial networks sharing content from the same brand (including the news outlets Focus and Tag24); networks with shared anti-establishment views; closely connected to that network, fan groups for the far-right party AfD; and others. AfD networks could clearly be traced back to specific AfD politicians who were administrators of apparently independent supporter groups; similar page names also appeared again and again on Facebook.
Such coordinated networks, and especially those affiliated with the AfD, were 32% more likely to share news sources that had been blacklisted by fact-checkers, including for instance RT or Epoch Times. The full report on this work is also available.