Longitudinal change in language behaviour during protests: A case study of Euromaidan in Ukraine.

In the last decade, online social media has become the primary platform for protesters to organize and express their agenda in various parts of the world. Nevertheless, scholars still debate whether online tools induce offline protests or facilitate them. Unfortunately, studies of protests often lack panel data and cannot address how particular users change their behaviour over time in line with the protest agenda. To this end, we analyze a new dataset of the Facebook page EuroMaydan that was explicitly created to facilitate the protest in Ukraine from November 2013 to February 2014. Moreover, our analysis follows this page even after the end of the protest till June 2014. In total, the dataset includes 26,631 posts and 1,470,593 comments that were generated by 124,790 users during and after the protests. We use this panel data to test how particular users switch between the two most popular languages of this page: Ukrainian and Russian. Previous studies have puzzled with language behaviour during the Ukrainian protest. Although researchers expected to see more Ukrainian language due to national mobilization, studies discovered that Ukrainian protesters frequently used Russian, especially after the end of the protest. A hypothesis was suggested that protesters use language strategically depending on circumstances (e.g., to maximize outreach). However, previous studies relied on aggregated data and did not explore the within-subject variation of language behaviour. Our study adds to this scholarship by adding a longitudinal analysis of language behaviour. Considering the broader contribution of this paper, we add to the literature on protests by validating previous findings derived from surveys that activists change their behaviour to reflect prior preferences and situational goals rather than modify their preferences.