The Commentariat: A Demographic, Ideological and Rhetorical Analysis of Twitter Power-Users

Luke Coughlan

Research output: ThesisDoctoral Thesis

98 Downloads (Pure)

Abstract

When political issues are discussed in the public sphere, whose opinions do we hear? This is a central democratic question, as those who commentate on political issues have a profound effect on public opinion (Zaller, 1992; Page, 1996). However, the way political commentary operates on social media has yet to be fully explored. This is a crucial omission, as political communication on social media platforms is dominated by a select few individuals (Bracciale et al., 2018). This subset of “power-users” – individual accounts who receive significant audience engagement – produce a large share of the political content that Twitter users are exposed to (Guo et al., 2020). Yet past literature has yet to answer three questions: 1) Who are Twitter’s political power-users from a professional and demographic perspective? 2) How extreme are power-users in terms of their issue opinions? And 3) To what extent do power-users influence the tone of political discussion on the platform? This thesis presents three empirical papers to investigate these questions, leveraging a dataset of 6.5 million tweets on Brexit, a highly salient issue which mobilised both elites and the public (Hobolt et al., 2018). The key contribution of this thesis is the identification of citizen opinion leaders – members of the public who possess a significant online political audience – as a new category of online political commentator. Citizen opinion leaders make up a large contingent of the most-shared accounts on Twitter (Paper #1) and play a central role in the spread of incivility in online political discussions (Paper #3). More broadly, this thesis provides evidence of how issue extremity and incivility on Twitter is driven by the platform’s political “commentariat”, a topic neglected by prior literature (Papers #2 and #3). Methodologically, it presents novel applications of text-scaling and time-series analysis to investigate such effects.
Original languageEnglish
QualificationPh.D.
Awarding Institution
  • Royal Holloway, University of London
Supervisors/Advisors
  • Smets, Kaat, Supervisor
  • Collignon, Sofia , Supervisor
Award date1 Apr 2024
Publication statusUnpublished - 2024

Keywords

  • social media
  • Twitter
  • politcal communication

Cite this