Unifying Models of Intergroup Hostility in Online Discourse

What changed

Researchers have introduced a novel empirical framework designed to unify and analyze six distinct theories of intergroup hostility commonly observed in online discourse. These theories, including boundary construction, threat construction, scapegoating, negative evaluation, dehumanization, and action orientation, were previously studied in isolation. By applying this unified model to a dataset of 2.86 million posts from TikTok, Truth Social, and Twitter/X during the 2024 U.S. presidential election, the study offers insights into the structural and temporal relationships between these mechanisms.

Why it matters for builders

This work provides a more cohesive understanding of how hostile rhetoric operates online, moving beyond single-label detection. For AI builders, this means a clearer empirical foundation for developing more sophisticated models capable of identifying, analyzing, and potentially moderating complex patterns of intergroup hostility. It bridges theoretical gaps in social science, offering a computational approach to a critical societal issue.

Practical impact

The study found that boundary construction and threat construction appear to anchor the system of intergroup hostility rhetoric. Temporally, a regular ordering was observed: boundary construction, derogation, and action orientation tend to appear early, followed by dehumanization and threat construction, with scapegoating emerging latest. This temporal sequencing offers valuable information for real-time analysis and intervention systems.

Caveats and source limits

The findings are based on an analysis of posts from a specific period (2024 U.S. presidential election) and three platforms (TikTok, Truth Social, Twitter/X). The study models existing theories within a common framework rather than proposing entirely new mechanisms. Further research may be needed to validate these findings across different contexts, platforms, and time periods.