They are more partisan, coordinated and determined to spread their messages virally through automated systems such as bots. This description captures the behaviour of pro-Republican far-right groups during the 2020 US presidential election on Twitter. In contrast, supporters of the Democrats appeared to be less coordinated overall and mostly engaged in harmless, fact-based communication. On the eve of the new US elections, this picture emerges from a study conducted by the University of Pisa and the CNR Institute of Informatics and Telematics published in the journal EPJ Data Science, of the Springer Nature group.
The research analysed a dataset of 260 million tweets and identified three main categories: moderate groups genuinely interested in the election debate, conspiracy groups spreading false information and controversial narratives, and foreign influence networks seeking to exploit the debate. Beyond the behaviour of right-wing and left-wing supporters, the results also showed that Twitter was only effective in countering the activity of some groups, while failing to counter other equally suspect groups.
“In a context of growing concern about the manipulation of public opinion through social media, our study offers an in-depth analysis of the dynamics on Twitter (X-hour) during the 2020 presidential election, says Marco Avvenuti (photo) Professor at the Department of Information Engineering of the University of Pisa.
“Coordinated online behaviour is a complex phenomenon that can have a significant impact on society and politics, making it essential to both analyse and understand it. As the 2024 elections approach, these findings provide critical insights into how to address threats to democracy and public communication, and help to ensure the integrity of online political debate.”
“Among these findings, we also noticed that some foreign communities were more active and determined in influencing the electoral debate, aiming to draw attention to the protests taking place in their countries," adds Serena Tardelli, researcher at the CNR Institute of Informatics and Telematics.