Political Attacks in 280 Characters or Less: A New Tool for the Automated Classification of Campaign Negativity on Social Media

Negativity in election campaign matters. To what extent can the content of social media posts provide a reliable indicator candidates' negativity? We introduce and critically assess an automated classification procedure that we trained to annotate more than 16,000 tweets candidates competing 2018 Senate Midterms. The algorithm is able identify presence political attacks (both general, specifically for character policy attacks) incivility. Due novel nature instrument, article discusses external convergent validity these measures. Results suggest classifications are measurements negativity. Triangulations with independent data show our automatic strongly associated experts’ perceptions candidates’ campaign. Furthermore, variations measures negativity be explained by theoretically relevant factors at candidate context levels (e.g., incumbency status gender); meaningful trends also found when replicating analysis using 2020 election, coded classifier developed 2018. implications such results coding discussed.