Do Word-Count Constraints Affect Populism and Verbal Violence? Evidence from Twitter Data in Italy

Our research hypothesis is that space restriction on social media communication creates fertile ground for aggressive, insulting and apodictic (AIA) speech thereby deteriorating the political debate and enhancing polarisation, violence and populism. We test our hypothesis by using as identification mechanism the decision of Twitter to double tweets’ length (from 140 to 280 characters) since November the 7th 2017. We use a machine learning classification approach and find with a discontinuity design diff-in-diff estimate that the event significantly and markedly reduced likes and retweets of aggressive, insulting and apodictic tweets in the political domain in the three-day interval around the policy change. Falsification tests show that our result is not spurious and the reduction in the share of aggressive, insulting and apodictic replies to a given political tweet after the character count change goes in support of our hypothesis. Our conclusion is that the Twitter reform has reduced attractiveness of aggressive speech thereby making the social media platform a less fertile ground for polarisation