A Cross-Platform Analysis of Far-right Online Communication: Studying Changes in Hate Speech Prevalence in the Aftermath of Terrorist Attacks

In late 2020, Europe was shocked by a series of terror attacks in Nice and Vienna. Such crisis events prove to be challenging for liberal democracies, as terror attacks aim at creating chaos. Terror paralyzes vast amounts of the general public by fostering a climate of shock, fear, and disorientation. In order to process these emotions, people post and connect with each other on social media for collective sense-making. At the same time, Islamist terror attacks increase public islamophobic sentiment and foster not only the promotion of xenophobia but also fuel radicalization dynamics (Fischer-Preßler, Schwemmer & Fischbach, 2019). Previous research finds that far-right actors actively take advantage of crisis events by trying to inject far-right narratives into public discourse with the aim to mobilize, recruit new members, and normalize extreme attitudes (Conway, Scrivens & McNair, 2019). As a result of comprehensive deletion actions by large platforms, far-right actors increasingly move to instant messaging platforms, like Telegram, for undisturbed communication within their communities. Such platforms function as safe havens for extreme or hateful ideologies and thereby support the development of fringe communities and possibly radicalization dynamics (Guhl, Ebner & Rau, 2020). We ask whether hateful content in the aftermath of terrorist attacks differs in its quantity and xenophobic nature. As an additional research question, a comparison with respect to quantity and severity of hateful content on instant-messenger spaces like Telegram and public social media sites on Facebook was performed. We address these research questions via manual quantitative content analysis of far-right communication on public Facebook pages (posts & comments) and Telegram groups. For a first preliminary study, we collected 26,456 Facebook posts and comments from eight popular far-right sites and 6,125 Telegram posts from five groups and actors, that were published prior and after the terror attacks (23.10. - 09.11.2020). Employing stratified random sampling, we manually coded 1,901 Facebook and 1,728 Telegram messages for severity, ranging from insults to call for violence, and type of hate speech. Preliminary results show a significant increase of anti-muslim hate speech in the aftermath of terror attacks. However, this increase is not accompanied by an overall increase in quantity or severity of hate speech when considering other types of hate speech. Contrary to previous research, which finds greater amounts of hateful content on fringe communities within 4chan or Gab (Zannettou et al., 2018), our cross-platform comparison finds more hate speech on public Facebook pages than in Telegram groups. A possible explanation might be that far-right individuals purposely communicate with hateful sentiment to provoke counter arguments and hence, receive more attention in public spaces. In consequence, hateful content may serve as a (radical) counterpart for mainstream collective sense-making narratives in the aftermath of terror attacks (Hohner, Schulze & Rieger, 2021). Additional sampling and analyses will be presented at the conference to further conclude on the theoretical mechanisms behind this finding.