From Variance to Invariance: Qualitative Content Analysis for Narrative Graph Annotation

Researchers developed a new framework for annotating economic narratives in news using directed acyclic graphs to represent causal relationships between events. The study focused on inflation narratives and introduced quality measures to reduce annotation errors, finding that lenient metrics overestimate reliability while locally-constrained representations improve consistency.

• →A new narrative graph annotation framework integrates qualitative content analysis principles to improve annotation quality for economic news narratives.
• →The framework represents inflation narratives as directed acyclic graphs where nodes are events and edges show causal relationships.
• →Research shows that overlap-based distance metrics overestimate annotation reliability compared to stricter measures.
• →Locally-constrained representations using one-hop neighbors significantly reduce annotation variability among human annotators.
• →The annotation framework and graph-based evaluation tools have been open-sourced for broader NLP research applications.