#Populism on twitter: Statistical analysis of the correlation between tweet popularity and “populist” discursive features

Abstract

Recent political events, such as the Brexit or Donald Trump's electoral success, have led to a proliferation of studies focusing on populism nature (Müller 2017; Mudde and Kaltwasser 2017). Part of the literature has also investigated communicative aspects of populism, highlighting how populists are benefitting from the use of social media (Bartlett 2014; Gerbaudo 2018). This research offers further insights on the subject by analyzing populist discourse on Twitter and exploring the correlation between the presence of linguistic features linked to populism, such as emotionalization, simplified rhetoric and intensified claims (Canovan 1999; Heinisch 2008), and tweet popularity. The use of linear mixed effects models revealed a positive correlation between the linguistic elements of interest and tweet popularity, not only in the populist sample, but also in the control group composed by establishment politicians. Surprisingly, reference tweets received more popularity than populist messages when the discursive features analyzed were present.