Which Topics Polarize Where? Mapping Polarization and Depolarization in News Comments

Research increasingly highlights the affective and destructive dynamics visible in everyday digital communication. However, digital trace research exploring related topics often remains concentrated on major social platforms (e.g., X/Twitter, Reddit) and WEIRD contexts. As a result, news forums and comment ecosystems are underexplored, despite their potential to intensify polarized views through disagreement (Kubin & von Sikorski, 2024). Using engagement data from Austrian online news forums (e.g., Der Standard), we examine polarization in article comments between January 2024 and December 2025, spanning different political orientations and media styles (e.g., left vs. right; quality vs. populist). By doing this, our primary goal is to identify which topics are most associated with high polarization-based language in these forums and whether polarization differs across news outlets. Furthermore, we aim to examine within-comment shifts in polarization to understand what depolarizes individuals within these comment threads. Expected outcomes include (1) a Polarization Index for comment and article-level data, (2) a ranked map of polarizing topics across outlets, (3) estimates of outlet differences in polarizing topics, and (4) evidence on which reply strategies/users are associated with de-escalation, which can potentially inform future interventions.