Shared Patterns: A Cross-Country Investigation of Users Reactions to Parties’ Facebook Posts During the 2019 European Parliament Election.

On Facebook, patterns of user engagement largely shape what types of political contents citizens can see on the platform. Therefore, one of the major goals of political actors’ Facebook communication is to produce content that hold the potential to trigger user engagement, and thereby increase their own visibility. There is a growing body of work that addresses the question of what types of political posts are effective in the triggering of engagement, but these efforts are generally limited to single-country investigations that decrease the ability to generalize findings. This work presented here is designed to fill this research gap. It does so by exploring the patterns of user engagement on a cross-country (Austria, France, Germany, UK, Hungary, Ireland, Italy, Poland, Romania, Spain, Sweden) sample of parties - 9406 Facebook posts from 66 parties - in the context of the 2019 European Parliamentary Election. The study will investigate how the forms, styles, topics and purposes of posts published by political parties affect the number of likes, negative reactions, comments and shares they trigger. In line with the literature, we hypothesize that posts containing visual elements, negativity, and personal details about politicians will trigger more reactions, comments and shares. Moreover, the study will particularly focus on the engagement-triggering effects of key topics of the election campaign such as immigration, climate change, labor and social issues, economic issues and domestic politics - as well as of populist communication. We hypothesize that immigration and climate issues were more popular among followers than more traditional campaign topics such as economy, social policy and domestic policy. Furthermore, the models include interaction terms to test if there is an issue ownership effect in user engagement in post-level, i.e. parties’ posts in their own issues trigger more likes, negative reactions, comments and shares. Finally, we also assume that posts containing populist elements such as elite-criticism, reference to the people or to dangerous others, will be more engaged with, irrespective of party ideology. These hypotheses are tested through statistical analysis using count models and taking into consideration the multilevel nature of the data.