Bots supplied nearly all negative sentiment in the Catalan referendum debate

Evidence

For example, one big social media study looked at over 4 million Twitter posts during the 2017 Catalan referendum – an unconstitutional referendum asking citizens whether they would prefer for Catalonia to become independent of Spain. In total, a little less than 25 per cent of the posts came from bots, but almost 40 per cent of the replies came from bots. A ‘sentiment analysis’ of the interactions and hashtags used by individuals revealed two clearly opposing camps: ‘Constitutionalists’ and ‘Independentists’ The independence camp revealed strong associations between terms such as ‘freedom’ and independence’, and ‘fight’ and ‘shame’, in relation to the Spanish government. What’s interesting is that the negative associations solely came from bot accounts. Moreover, although bots appear to be on the periphery of the social network, they gained influence by strategically targeting human influencer accounts, who were more centrally positioned within the network of independentists. — Sander van der Linden, Foolproof, 732

Categorization

• XREF: connects to: Bots manufactured roughly 20 percent of pre-election political conversation — Russell Muirhead and Nancy L. Rosenblum, A Lot of People Are Saying: The New…, 81