Politicians’ Communication on Facebook and Followers’ Engagement

• In Event: What Makes It Go Viral?

Abstract

Across Western democracies social media are widely used by politicians to connect with the electorate, and they have become a central political battle field in politicians’ fight for attention and citizen engagement. Yet, not all communications from politicians succeed in engaging citizens’ attention and go viral on social media. Still, little is known about which type of political elite communications that trigger citizen engagement on social media in the form of for example “likes,” “shares,” and “comments” This paper therefore asks: Which type of political elite communication on social media that triggers online response from citizens?
Prior studies emphasize that a central reason why people opt in to follow a politician on social media is that they are interested in politics and want information from other sources than the traditional news media orientation (e.g. Krapoulouros 2017). On this basis, we expect that messages that focus on major political events and outline politician’ issue opinions will receive more likes, comments and shares the followers than non-political content. Importantly, research on the negativity bias suggests that messages with a negative tone trigger more attention and receive more comments and shares from followers in online networks than positive messages (Bene 2017, Heiss, Schuck & Matthes 2018). We predict that the effect of negative sentiment on follower engagement and information sharing on social media should be stronger for political content than non-political content as politics is fundamentally competitive and conflict-driven (e.g. Schattschneider 1942). and the party antagonism drive people to focus on problems (e.g. Rosenblum 2011).
We test our predictions using a unique data set of all Facebook text posts from the public Facebook pages of Danish members of parliament (MPs) in 2015 (n = 124,223) and audience responses to these posts in the period 2008 to 2018. We use human coding, dictionary-based sentiment analyses and machine learning to code the content of the posts. Our results support the predictions and we discuss the implications for democratic elite-citizen dynamics on social media.

Authors

• Lene Aarøe, Aarhus University

• Morten Brænder, Political Science, Aarhus University

• Matthew W. Loftis, Aarhus University