Social bots in election campaigns: theoretical, empirical, and methodological implications

Abstract

Social bots mimic and potentially manipulate humans and their behaviours in social networks. The public sphere might be especially vulnerable to their impacts, which is why we first discuss their potential influence on the public sphere from a theoretical perspective. From an empirical perspective, we analyzed Twitter followers of seven German parties before (N = 638,674) and during (N = 838,026) the 2017 electoral campaigns regarding bot prevalence and activities. The results revealed that the share of social bots increased from 7.1% before to 9.9% during the election campaigns. The percentage of active social bots remained roughly the same. An analysis of the content distributed by both the most popular and the most active bots showed that they disseminate few political hashtags, and that almost none referred to German politics. We discuss the results against the background of normative traditions of public sphere theories and address the methodological challenges bots pose in political communication.