APS

30th APS Annual Convention · 2018

Tweeting Anger Predicts County-Level Results of the 2016 United States Presidential Election

San Francisco, CA · May 2018

Poster · Social

  • Katharina Bernecker
    Leibniz-Institut für Wissensmedien Tübingen
  • Michael Wenzler
    Leibniz-Institut für Wissensmedien Tübingen
  • Kai Sassenberg
    Leibniz-Institut für Wissensmedien Tübingen

Abstract

We used a sample of 148 million tweets to predict the results of the 2016 US Presidential Election. Donald Trump received more support in counties, where people tweeted more anger and negative emotions, even when various county characteristics (e.g., education) and conservative vote choice in preceding presidential elections were controlled.

Big Data

← Poster Session X