The mass public's science literacy and co-production during the COVID-19 pandemic: empirical evidence from 140 cities in China

The COVID-19 pandemic highlighted the crucial role of public participation (co-production) in pandemic response. Effective public health policies depend on widespread voluntary cooperation. While existing research has explored individual factors influencing co-production, the role of science literacy remains understudied. This study addresses this gap by empirically examining the relationship between public science literacy and co-production during the COVID-19 pandemic in China. The study focuses on China due to the extensive public collaboration observed during the pandemic and the relative lack of ideological biases that might confound the relationship between science literacy and cooperation. The authors define co-production in this context as the collective efforts of the public and government in fighting the novel coronavirus. The study uses data from a national survey on civic science literacy and Baidu Index (search data) to measure co-production, controlling for various environmental and individual factors. The research hypotheses are: (H1) Public science literacy significantly promotes co-production; (H2) Regional education levels positively moderate the effect of science literacy on co-production; and (H3) Local government capacity positively moderates the effect of science literacy on co-production.

The literature generally agrees that trust in science, scientific spirit, and scientific knowledge facilitate effective responses to complex global challenges. Higher science literacy is associated with better understanding of pandemic prevention policies and greater adherence to measures like mask-wearing and vaccination. Conversely, anti-intellectualism and science skepticism hinder compliance. However, existing studies on science literacy and co-production during COVID-19 have limitations. These include potentially one-sided indicators of science literacy, limited sample scope, and reliance on subjective measures of co-production behaviors. This study aims to address these limitations using comprehensive national-level data from China and an objective measure of co-production based on online search behavior.

The study employs an econometric approach using Stata 17.0. The primary model uses ordinary least squares (OLS) regression with public co-production (measured by Baidu Index search volume for COVID-19-related keywords) as the dependent variable and science literacy (percentage of citizens with science literacy in a city) as the key independent variable. Control variables include environmental factors (GRP, income, technology level, number of foundations, government internet transparency, digital government development) and individual-level factors (government trust, social trust, social justice). To address potential endogeneity, a two-stage least squares (2SLS) model is used with the ratio of R&D personnel to population as an instrumental variable for science literacy. Moderating effects of regional education level (proportion of middle school students) and local government capacity (number of discredited people at the provincial level) are examined by including interaction terms in the model. Heterogeneity analysis is conducted across different pandemic stages, regions, and city sizes using subsample regressions. The data encompasses 140 cities across China, representing about half of the country's population. The study uses data from the 2020 National Survey of Civic Scientific Literacy in China and Baidu Index.

Baseline OLS regression shows a significant positive relationship between science literacy and co-production (β = 0.142, p<0.01). For every 1% increase in science literacy, co-production increases by 14.2%. The 2SLS model, addressing endogeneity, strengthens this relationship (β = 0.425, p<0.01), suggesting the OLS model underestimated the effect. The moderating effects analysis reveals that higher regional education levels and stronger local government capacity significantly amplify the positive impact of science literacy on co-production. Heterogeneity analysis shows that the effect of science literacy varies across pandemic stages, being particularly strong during the initial and final stages. Regional differences also exist, with the central and western regions showing stronger effects than the east. City size also plays a role, with the effect being weaker in larger cities. Robustness checks using different sample selections confirm the key findings.

The findings demonstrate the crucial role of public science literacy in fostering voluntary compliance with pandemic prevention measures. This contrasts with approaches relying solely on authoritarian mandates. The positive moderating effects of education and government capacity highlight the importance of broader societal factors in amplifying the benefits of science literacy. These results align with previous research showing the negative impacts of anti-intellectualism and science skepticism on pandemic responses, but offer a complementary perspective by focusing on the positive influence of science literacy. The heterogeneity findings suggest that targeted interventions may be needed to tailor communication and education strategies to specific contexts.

This study provides strong evidence supporting the positive impact of public science literacy on co-production during a pandemic. The findings underscore the importance of investing in science education and communication, alongside building strong local governance capacity, to effectively engage the public in crisis response. Future research could explore the dynamics between science literacy, trust in institutions, and co-production using longitudinal data and more sophisticated modeling techniques to understand nuances in the relationships revealed here.

The study's cross-sectional design limits causal inferences. While the instrumental variable approach helps address endogeneity, other unobserved factors could still influence the relationship. The reliance on Baidu Index as a proxy for co-production, though objective, might not capture all dimensions of public participation. The sample, while large, is limited to cities in China and might not fully generalize to other contexts. Future research should explore these limitations using longitudinal data and diverse measures of co-production.