Beyond the Battlefield: Framing Analysis of Media Coverage in Conflict Reporting

TL;DR

This study employs computational methods and large language models to analyze media framing in conflict reporting, revealing biases and differences in coverage across regions during the Israel-Palestine war.

Contribution

It introduces a novel computational approach combining frame semantics and large language models to analyze conflict framing in news articles.

Findings

Higher focus on war framing over peace framing

Regional differences in reporting biases

Identification of media biases in conflict coverage

Abstract

Framing used by news media, especially in times of conflict, can have substantial impact on readers' opinion, potentially aggravating the conflict itself. Current studies on the topic of conflict framing have limited insights due to their qualitative nature or only look at surface level generic frames without going deeper. In this work, we identify indicators of war and peace journalism, as outlined by prior work in conflict studies, in a corpus of news articles reporting on the Israel-Palestine war. For our analysis, we use computational approaches, using a combination of frame semantics and large language models to identify both communicative framing and its connection to linguistic framing. Our analysis reveals a higher focus on war based reporting rather than peace based. We also show substantial differences in reporting across the US, UK, and Middle Eastern news outlets in framing who the assailant and victims of the conflict are, surfacing biases within the media.