Document Type
Thesis
First Faculty Advisor
Suhong Li
Keywords
political polarization sentiment analysis twitter
Publisher
Bryant University
Rights Management
CC-BY
Abstract
Amidst politically strained times, one might wonder what has cause such an exaggerated gap between the views of democrats and republicans. For years, research has suggested the US’s voting population is becoming increasingly politically polarized, with one of the causes being social media. This study's purpose is to understand more about the role that social media plays in the polarization of parties in the US. The study is comprised of the analysis of over 3,000,000 tweets from 9/22/2020 through 11/10/2020 that mention or are written by senate and presidential candidates. Natural language processing, network graphing, and sentiment analyses were utilized to draw conclusions about the polarity of political discussion in the 2020 election. Based on the analysis, there are two distinct “echo chambers”—one republican and one democrat – that are very rarely interacting with one another. There are also spikes in sentiment around election day, suggesting polarity increases close to this day. Finally, popular negative topics gain rapid traction on twitter compared to more neutral topics, showing their infectious nature. These findings will contribute to the existing body of research surrounding the effect that social media has on the polarization of parties in the US.
Included in
American Politics Commons, Data Science Commons, Other Computer Sciences Commons, Political History Commons
