Kulichkina, A., Balluff, P., Righetti, N., & Waldherr, A. (2026). Connective action and digital repression during China’s COVID-19 protests: a computational analysis of multilingual coordinated activity on Twitter. EPJ Data Science, 1–25. https://doi.org/10.1140/epjds/s13688-026-00637-2
Summary
This paper examines coordinated activity on Twitter surrounding China’s 2022 COVID-19 protests, arguing that social media in authoritarian contexts operates as dual-use infrastructure — simultaneously enabling grassroots protest mobilization and facilitating state-aligned repression. Drawing on frameworks of connective action and digital repression, the authors use a coordination detection algorithm to identify accounts engaged in coordinated communication and characterize their thematic, temporal, and linguistic signatures. The central insight is that contentious politics online is not a one-sided affair: protest coordination and repressive counter-narratives coexist and compete within the same platform.
Key Contributions
- Empirical evidence on the scale and structure of coordinated communication during a major protest event in an authoritarian setting.
- Demonstrates application of coordination detection methods to multilingual social media data on Chinese contentious politics.
- Bridges connective action theory and digital repression research through a case study of competing coordinated narratives.
Methods
- Computational analysis of Twitter data related to China’s COVID-19 protests.
- Application of a coordination detection algorithm to surface accounts exhibiting coordinated behavior.
- Multilingual analysis of linguistic patterns across coordinated activity.
- Examination of thematic content and temporal dynamics of the identified accounts.
Findings
- 13,557 Twitter accounts were identified as involved in coordinated activity around the protests.
- Coordinated communication displayed identifiable and distinct patterns across themes, time, and language (specifics beyond the abstract not available).
- Coordination served both mobilizing and repressive functions, consistent with the dual-use framing.
Connections
This work sits within the coordinated-behavior detection literature; its algorithmic approach connects to methodological efforts to identify coordinated link-sharing and inauthentic activity such as Giglietto2020-9d8acdd7, Giglietto2022-0e951ac5, and Giglietto2023-fa71a001, as well as broader coordination-detection studies like Luceri2025-tr and Minici2024-tf. Its focus on state-aligned repression and protest in an authoritarian context resonates with research on platform politics in similar settings, notably Kuznetsova2025-nu and Bosch2024-hj.