Causal Language as a Tool for Structural Intervention: Experimental Manipulation of Message Framing to Disrupt Echo Chambers

Shi & Morstatter (2024) find that messages with causal language spread further and more effectively bridge out-groups, suggesting a powerful lever for mitigating polarization. This idea proposes a series of field or lab experiments in which message framing is manipulated—injecting or removing causal language (e.g., “X causes Y” vs. “X is associated with Y”)—in targeted parts of a real or simulated social network. Outcomes would include both diffusion reach and downstream attitude change, carefully measured with causal inference techniques to separate framing effects from network structure effects. The study could also explore how these effects interact with user attributes (e.g., prior beliefs, centrality) and echo chamber strength (as characterized by Yu et al., 2025). This goes beyond observational studies by directly testing a new form of “structural intervention”—message design—opening up a fresh practical tool for platforms, educators, or public health communicators seeking to promote information diversity.

References:

1. Breaking The Loop: Causal Learning To Mitigate Echo Chambers In Social Networks. Dianer Yu, Qian Li, Huan Huo, Guangdong Xu (2025). ACM Transactions on Information Systems.
2. The Diffusion of Causal Language in Social Networks. Zhuoyu Shi, Fred Morstatter (2024). International Conference on Web and Social Media.