Applying Topic Modelling to Party Discourse: An Exploration of the Italian Case in 2013–2019

Abstract

During the last few years, external crises and endogenous weaknesses have combined to plunge the Italian political system into generalised instability. In particular, the major political parties have experienced rapidly turning tides in a context of intensified electoral volatility. This explorative article sets out to get an insight into the discursive struggles that have pitted these parties one against another, undergirding the ebb and flow in their respective mass support and revolving around the ways of communicating political change. To that end, I collect data from the official Facebook pages of the four main Italian parties, downloading posts they published in the period 2013–2019 via the Netvizz application, and I analyse the four corresponding textual corpora through the technique of Topic Modelling. On such bases, the article finds the overall configuration of the political discourse of Italian parties to be aptly described by a model comprising 16 topics, equally divided into ‘partisan’ and ‘cross-cutting’ ones, with the former having a slight edge in terms of diffusion. The four parties differ among themselves by the topics they focus on and by the quantity of topics they choose to include sizably in their streams of communication.