Policy feedback in crowd-safety crises: a dynamic topic modeling approach to South Korea’s Post-Itaewon crisis

In October 2022, a tragic crowd crush in Itaewon, South Korea, triggered intense public outrage over the government’s unpreparedness. This study examines how public discourse and policy actions interacted over time, using policy feedback theory to analyze how government responses influenced trust, engagement, and meaning making. Online data were collected from Korea’s largest platform, Naver, and dynamic topic modeling was applied to identify evolving themes in citizen narratives. These trends were then aligned with the timeline of policy measures, ranging from symbolic gestures to more structured reforms. Results reveal a dynamic feedback loop: early policies were perceived as symbolic gestures responding to political pressure, while subsequent reforms that gained greater public visibility and addressed prevailing concerns gradually redirected discourse toward future-oriented preparedness. However, a resurgence of concern over site-specific risks in late 2024 indicated that technocratic solutions alone may not fully restore trust or closure. By tracing shifts in online discourse alongside government actions, this study highlights the interpretive mechanisms through which citizens assess post-crisis governance. Findings offer insight into when and how policy responsiveness supports democratic legitimacy, and underscore the importance of visible, emotionally resonant, and experientially credible reforms in maintaining public confidence after disaster.