Assessing the Depth and Width of Deliberative Discussions Online

The second day at the ACSPRI 2024 conference dawns with a session on social network mapping, and starts with a paper by our wonderful conference chair Rob Ackland. This presents work on an international collaboration around technology and political communication, with a particular focus on social bots. This explores especially the potential for such bots to connect people with different ideas online, with the aim to improve public discourse.

This requires us, in the first place, to understand where discussion and deliberation are occurring in online spaces: deliberative conversations require both diversity (or representation of different ideas) and argumentation (a genuine discussion of those ideas). These two factors can be understood as the width and the depth of online discussion, respectively.

This project explored this, for instance, for the first of the 2020 US presidential election debates on Twitter, assessing the reply trees stemming from original tweets with the #DebateNight hashtag. (This kind of work is now no longer possible, thanks to the enshittification of Xitter as a platform under Elon Musk.) These reply trees can then be assessed for their depth and width, and genuinely deliberative conversations singled out.

Further work beyond this study must then also take into account what is actually said in these posts, beyond an assessment of the network structure alone – there may be incivility or conflict in those replies, for instance.