Researcher ORCID Identifier

0000-0002-9660-8832

Graduation Year

2025

Date of Submission

4-2025

Document Type

Open Access Senior Thesis

Degree Name

Bachelor of Arts

Department

Economics

Reader 1

Cameron Shelton

Terms of Use & License Information

Terms of Use for work posted in Scholarship@Claremont.

Rights Information

© 2925 Nolan D Windham

Abstract

This thesis empirically investigates the relationship between exposure to time-varying partisan news content and individual policy attitudes. Combining individual-level Cooperative Election Study data (2020–2023, N≈68,000 for seven policy outcomes) with daily Fox News and MSNBC transcripts, I analyze how fluctuations in media narratives associate with public opinion. Leveraging large language models, I classify all relevant news segments by topic and ideological stance (liberal/conservative), generating daily time series of content volume and slant for each channel and policy issue. These series construct respondent-specific exposure measures based on self-reported viewership and 7-day rolling aggregates of media content preceding the survey interview date. OLS regression models predict standardized policy attitudes using interaction terms between viewership and media aggregates (Net Tone; or Volume and Slant simultaneously), controlling for demographics, ideology, party identification, and county and year fixed effects. The analysis finds a robust association between the ideological slant of recent media exposure and policy attitudes. Controlling for content volume and fixed effects, exposure to more liberal-slanted coverage in the preceding week significantly associates with holding more liberal views across nearly all policy domains (most p<0.001). For instance, a one-unit increase in the normalized liberal slant (-1 to +1) of Fox News' coverage of assault weapons over a week correlates with a 0.146 (SE=0.012) point increase in viewer support for a ban. This slant effect generally dominates the smaller and less consistent associations found for content volume. These findings demonstrate a strong correlation between how partisan outlets frame issues and viewer opinions, distinct from channel selection, highlighting the value of granular content analysis for understanding media's role in shaping policy views.

Download

Share

COinS