Framing Conflict

Media framings of civil wars have long shaped public opinion, thereby influencing international responses and intervention decisions. While existing studies have relied primarily on qualitative evidence or focused on a single war dyad, less is known about how specific media framings affect civil conflicts more broadly using cross-national data. To address this gap, we introduce the Conflict Framing Dataset, a new resource that enables systematic, large-scale analysis of how international media framing influences civil war dynamics. By utilizing ConfliBERT and fine-tuning the model, we identify dyads in news articles and classify them according to their dominant framing: humanitarian-focused versus politically-focused coverage. This paper outlines our methodology for constructing and validating the dataset and classifier, and discusses the potential applications of this dataset based on the classification results. By quantifying media framing at scale, our approach opens new avenues for research on the interaction between global media narratives and state behavior in civil war contexts.