Natural Language Politics

Open Access
- Author:
- Burnham, Michael
- Graduate Program:
- Political Science
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- June 10, 2024
- Committee Members:
- Lee Banaszak, Program Head/Chair
Burt Monroe, Special Member
Michael Nelson, Chair & Dissertation Advisor
Sarah Rajtmajer, Outside Unit & Field Member
Suzanna Linn, Major Field Member
Kevin Munger, Major Field Member
Bruce Desmarais, Major Field Member - Keywords:
- Natural Language Processing
Text as Data
Congress
Legislative Effectiveness
Affective Polarization
Ideology - Abstract:
- This dissertation explores the how recent advancements in text analysis methods can be applied to political research. In the first chapter, I demonstrate how language models can facilitate opinion mining at a much larger scale than was previously possible, and guide researchers through best practices in employing these methods. In the second chapter, I build on these methods by presenting a method of estimating ideology from text that is based on expressed opinions rather than the choice of words. Finally, in chapter three I demonstrate how these methods enable researchers to ask and answer questions about politics that may have been infeasible previously. The chapter uses text analysis to present the first scaling method for affective polarization among members of congress, and then uses this measure to test how affective polarization influences legislative effectiveness.