Leaving messages as coproduction: impact of government COVID-19 non-pharmaceutical interventions on citizens' online participation in China

The COVID-19 pandemic necessitated the implementation of various non-pharmaceutical interventions (NPIs) globally, including curfews, travel restrictions, school closures, and lockdowns. While the macro-level impacts of NPIs on the environment and economy have been extensively studied, their micro-level socio-political effects remain less understood. This study addresses this gap by focusing on the impact of NPIs on citizens' online participation in China during 2020. The increased social distancing and self-isolation imposed by NPIs significantly altered people's lives, impacting their mental health and daily routines. However, the shift in government-citizen interaction to online platforms, necessitated by lockdowns, presents a unique opportunity to analyze the effect of NPIs on online political engagement. Timely and effective communication between governments and citizens is crucial during pandemics. Online interactions provided an essential channel for information exchange and access to authoritative information, especially during periods of isolation and reduced traditional social interaction. This online interaction with the government is a critical form of online participation, offering a more convenient channel for engagement, especially for traditionally marginalized groups. Existing research on online political participation focuses on socioeconomic status, education, social networks, digital skills, and government media use patterns. This study adds a unique dimension by investigating the impact of exogenous shocks, such as the COVID-19 pandemic, on online participation patterns. The study focuses on China for two reasons: China's early and widespread adoption of NPIs in 2020 provides a suitable sample; and it addresses the research gap of studying online participation in developing countries, especially in a unique political and cultural context like China.

Existing literature highlights the extensive macro-level consequences of NPIs, including environmental improvements and economic downturns. Micro-level research primarily focuses on the behavioral and personal impacts of NPIs on individuals, such as mental health, sleep patterns, and dietary habits. The literature on online political participation emphasizes the role of socioeconomic factors, digital literacy, and government information strategies. However, limited research exists on the effects of exogenous shocks, like the COVID-19 pandemic, on online participation patterns, particularly in developing countries.

This study utilizes a difference-in-differences (DiD) model and a dataset comprising daily messages received by City Party Secretaries on the Local Leadership Message Board (LLMB) from January 1 to December 31, 2020. The LLMB is a nationwide online platform facilitating citizen-government interaction. Data was collected using a Python crawler. Cities implementing strict NPIs (prohibition of congregation, intercity and intra-city travel restrictions) formed the treatment group, while cities without these measures formed the control group. The DiD model compares the change in online participation (number of messages) between the treatment and control groups before and after NPI implementation, controlling for the severity of local COVID-19 outbreaks. To mitigate potential biases associated with DiD, the study also employed the interactive fixed effects counterfactual estimators (IFEct estimators), which construct counterfactuals for treated cities using data from control cities. The study performs robustness checks, including placebo tests, winsorization of extreme values, and analyses using various subsamples (first wave of pandemic, pre-Spring Festival period, excluding Spring Festival). Heterogeneity analysis explores differences in treatment effects across cities based on economic conditions (GDP per capita), telecommunication infrastructure (mobile phone subscribers), and educational levels (college graduates). Mechanism analysis examines the types of messages (seeking help, consulting, providing suggestions, expressing thanks, complaining) to understand the motivations behind increased online participation.

The DiD model revealed that the implementation of NPIs significantly increased citizens' online participation, with a 0.166 increase in the number of daily messages to City Party Secretaries after controlling for various factors (Model 2, Table 2). This effect remained robust in various robustness checks. The event study design validated the parallel trend assumption, showing a significant increase in the treatment group only after NPI implementation (Fig. 2). The IFEct estimator corroborated the DiD results, showing a 0.743 increase in messages compared to counterfactual estimates (Model 6, Table 4). Placebo tests confirmed the validity of the IFEct estimator (Table 3). Heterogeneity analysis (Table 8) showed that the effects of NPIs were significantly larger in cities with better economic conditions, telecommunication foundations, and higher education levels. Mechanism analysis (Table 9) indicated two driving mechanisms: increased demand for help (seeking help, expressing thanks) and active coproduction behaviors (providing suggestions). The analysis of message content (Fig. 4 and Supplementary Table 3) further supported these findings, revealing an increase in messages related to transportation, healthcare, and administration during NPI implementation. The temporal trend analysis (Fig. 1) suggests that the peak effect of the NPIs on online participation occurred around three weeks after implementation.

The findings demonstrate a significant and robust impact of NPIs on citizens' online participation in China. The results address the research question by providing strong evidence for a causal link between NPIs and increased online engagement. The significant increase in online participation during NPIs highlights the importance of robust digital infrastructure and citizen digital literacy for effective government communication during crises. The two mechanisms identified—increased demand for help and coproduction—offer valuable insights for future crisis management strategies. The time-bound nature of the peak effect underscores the need for timely adjustments of NPIs to balance pandemic control with minimizing adverse social and economic consequences. The findings challenge the traditional view of online and offline participation as mutually exclusive entities, suggesting a complementary relationship.

This study demonstrates a significant and robust causal link between COVID-19 NPIs and increased citizens' online participation in China. The findings highlight the importance of considering the socio-political effects of NPIs in pandemic planning and response. Future research should explore other online platforms, incorporate individual-level data, and examine the long-term sustainability of this behavioral change. Moreover, understanding the composition of online participation (e.g., coproduction vs. mere demand) is crucial for optimizing government responses to future crises.

The study relies on data from the LLMB, which may not capture all forms of online citizen-government interaction. The analysis focuses on China's unique political and technological context, limiting generalizability to other countries. While robustness checks were conducted, the complex interplay of factors during the pandemic may still influence the results. Further individual-level data is needed to analyze the variations in online participation across different demographic groups and to explore the long-term effects once the NPIs are lifted.