Bridging or Bonding? Measures of Topic Centrality for Online Political Engagement

Cameron Raymond


The advent of social media has enabled political parties to engage with the broader populous in new and unforeseen ways. The ability to bypass the traditional mediating forces of mass media allows for an unfiltered promotion of policy, ideology and party stances. This is specifically interesting in Canada’s political system which has historically been defined by large brokerage parties, who act as instruments of accommodation, taking a non-ideological stance in an attempt to maximize the number of different demographics willing to vote for them. In order to win a diverse range of electoral districts, these “big tent” parties try to appeal to various political persuasions. Political adverts, and policy have traditionally been the conduits through which brokerage parties attempt to accommodate different ideologies, but social media allows for a direct, granular approach to political messaging which is completely novel. This has inherent implications for our conception of political engagement - and reasoning over this large amount of data requires significant computational ability. Social networks, such as Twitter, are also easily represented as graphs. Therefore, as political strategy becomes increasingly digital, the use of graph theory can potentially illustrate how large brokerage parties organize and to what degree they are effective in reaching out to different political persuasions. The contribution of this piece of research is a large scale, robust, mathematical process - using topic modelling and measures of graph centrality - for evaluating how political parties distribute information and how users interact with these messages.

← Schedule