Facebook election advertising:dangerous for democracy or politics as usual? The case of the 2017 UK general election

Recent years have seen increasing concern about the potential of Facebook advertisements to microtarget small groups or individuals in the electorate with specific messages during campaigns and the impact such practices might have on democratic institutions. This article examines Facebook advertising during the 2017 UK general election using a dataset created by the civil society organization Who Targets Me. We use content analysis to code negativity and political topics to identify if certain messages disproportionately appear in the feeds of particular voters. Our empirical findings suggest that, while there is evidence of broad demographic targeting occurring, the single most important factor in what appears in a user’s Facebook feed is the point in the election campaign they are accessing Facebook. Surprisingly, given the UK’s first-past-the-post election system, we find limited evidence of geographical targeting. We situate these findings in the political context of the 2017 UK General Election, an unexpected “snap election” which saw large opinion poll movements during the campaign, and was called in the period of political dislocation following the UK’s referendum on European Union membership in 2016. As such, we argue that the election remains an important case study for understanding the interaction between the affordances of Facebook advertising and the political context.