Seat Security and Tweet Extremity: An Analysis of Congress Members’ Tweets
Open Access
- Author:
- Cohen, Emma
- Area of Honors:
- Political Science
- Degree:
- Bachelor of Science
- Document Type:
- Thesis
- Thesis Supervisors:
- Kevin M Munger, Thesis Supervisor
Michael Barth Berkman, Thesis Honors Advisor - Keywords:
- politics
social media
twitter
extreme
congress members
polarization - Abstract:
- Why is it that politicians have different tweeting approaches? What influences politicians to use their social media to enforce polarization? This study seeks to understand what factors drive extremity on social media using over 500,000 tweets from members of the 117th Congress. I predict that members of Congress who are not confident in their ability to be reelected will seek to appeal to more of the electorate through moderate tweeting and that members of Congress who are secure in their seat have the option to be extreme in their tweeting. Secure Congress members will only take this option to be more extreme when they have incentives to do so, such as personal beliefs, donor influences, or primary influences. After implementing a dictionary-based machine coding method to construct extremity scores for each member of Congress and regressing this on various potential influences, I find general support for my theory. Seat security is positive and significant across all models, demonstrating that Congress members who are less secure in their ability to get reelected are less extreme on Twitter, and Congress members with more seat security are more extreme. The interaction between seat security and donor reliance is also positive and significant, which provides support that high reliance on partisan donors increases extremity on Twitter when members of Congress are relatively secure. Additionally, I find that Republicans are much more likely to tweet extreme content than Democrats. These central findings are evidence that politicians are responsive to voters and donors even on social media and that the median voter theorem applies to Twitter behavior.