Abstract
This research-in-progress presents a model that decouples being seen from shaping participation in social media debates. Applied to Reddit’s communities, the approach constructs a directed user-interaction network from reply chains, estimates user-level activation tendencies with topic-conditioned sentiment, and evaluates seed users through systematic cascade experiments. Evidence from a tightly knit subnetwork indicates that fusing content signals with interaction structure pinpoints initiators associated with broader, more persistent cascades. The work offers a reusable blueprint for engagement-oriented influence mapping and actionable insights for researchers and practitioners in knowledge and information systems.
Recommended Citation
Belik, Ivan and Utochkin, Denise, "From Sentiment to Spread: Enhancing Influence Mapping in
Online Communities" (2025). ACIS 2025 Proceedings. 7.
https://aisel.aisnet.org/acis2025/7