Sentiment Analysis of Emotional Framing in Political Media

Abstract

Affective polarization, a persistent challenge in today's democracy, often overshadows the subtle yet profound influence of emotional rhetoric in shaping public opinion and reinforcing partisan divides. This research delves into the critical role of emotional framing in exacerbating this polarization within the U.S. media landscape, particularly across election cycles. This research trials a novel natural language processing algorithm, to measure and classify the emotional intensity of news content. This study, encompassing a vast collection of articles sourced from MediaCloud's sorted collections, uncovers significant variations in media's emotional intensity. We observed distinct partisan strategies in how emotional language is deployed, patterns that strikingly align with established theories of political psychology. Furthermore, our findings challenge the common assumption that emotional rhetoric intensity is confined to election seasons, revealing that such framing is deeply embedded in today's information-systems.