Leadership signals most strongly predict trust and COVID-19 compliance willingness

A randomized survey experiment involving 1,593 participants in China during the zero-COVID period found that perceptions of government leadership had the strongest effect on government trust. Trust, in turn, was associated with willingness to comply with protective behaviors, while transparency retained a smaller direct relationship with trust. Accountability and adaptability were no longer significant predictors once leadership and transparency were included in the model.

The study tested how different reputational signals embedded in crisis-communication messages may relate to public-health compliance. Drawing on Bureaucratic Reputation Theory, the researchers organized government reputation into four dimensions: leadership, accountability, transparency and adaptability. These dimensions were placed along two theoretical axes, competence–integrity and authority–responsiveness. The framework was designed to distinguish the psychological attributes that people may use when deciding whether to trust authorities and follow health guidance.

Participants viewed simulated crisis-communication messages containing different reputation signals. Manipulation checks showed that each signal produced its largest increase in the targeted perception dimension: leadership, accountability, transparency or adaptability. The experiment therefore demonstrated that the messages activated the specific reputation attribute they were intended to emphasize, allowing the researchers to compare how those attributes related to trust and compliance willingness.

When all four reputation perceptions were entered simultaneously into a structural equation model, leadership emerged as the dominant predictor of government trust. Transparency retained a modest direct path to trust. By contrast, accountability and adaptability lost their predictive power after leadership and transparency were accounted for, indicating that they did not explain additional variation in trust within the model. These results concern the relative contribution of the measured perceptions in this sample, rather than proving that accountability or adaptability is irrelevant in every crisis or institutional setting.

A nested model comparison further suggested that trust may fully mediate the association between reputation perceptions and compliance willingness. In statistical terms, the pattern was consistent with reputation perceptions operating through government trust, while direct paths from the four reputation dimensions to compliance willingness were not significant. The proposed mechanism is therefore leadership-anchored: perceptions of leadership were most strongly connected to trust, and trust was linked to willingness to comply with health measures.

The relationship between trust and compliance varied with the personal cost of the requested behavior. Trust had a stronger effect on willingness to perform high-cost actions, such as voluntary quarantine, than on willingness to perform low-cost actions, such as mask-wearing. The finding does not mean trust was unrelated to mask-wearing; rather, the observed association was stronger when compliance required a greater sacrifice. The results suggest that trust cannot be treated as a uniform driver of behavior across different levels of behavioral burden.

For crisis communication, the findings favor carefully distinguishing among reputational attributes rather than treating positive government signals as interchangeable. Leadership had the clearest relationship with trust, while transparency made a smaller independent contribution in the model. Accountability and adaptability were measured successfully, but they did not retain independent predictive relationships with trust after the other two dimensions were considered. The study thus identifies which attributes were most consequential within its experimental and analytical setting, without establishing a universal message strategy.

Several limitations define the scope of the findings. This was a randomized survey experiment measuring perceptions, trust and stated compliance willingness during China’s zero-COVID period, not a study of actual protective behavior or infection outcomes. The sample and institutional context also limit generalization to other populations, governance systems and public-health emergencies. The authors describe the results as a cost-contingent framework and identify whether the same trust mechanism operates elsewhere as a question for future research.

Within those boundaries, the experiment provides a technical account of how reputation signals may be connected to public-health compliance. Leadership was the strongest correlate of trust, transparency added a modest direct relationship, and trust was the principal pathway linking reputation perceptions to willingness to comply. The cost analysis adds an important qualification: government trust appeared most influential when people faced a more burdensome protective action. The evidence supports a leadership-centered, cost-sensitive interpretation of compliance willingness in this Chinese zero-COVID sample, rather than a general prescription or a demonstrated real-world health outcome.

Subject of Research: Government reputation signals, public trust and COVID-19 compliance willingness

Article Title: Crisis communication strategies, government reputation signals, and health compliance willingness during COVID-19: a randomized experiment in the context of China’s zero-COVID policy

Article References: Deng, W., Mao, Z., & Yang, Y. (2026). Crisis communication strategies, government reputation signals, and health compliance willingness during COVID-19: a randomized experiment in the context of China’s zero-COVID policy. BMC Public Health. https://doi.org/10.1186/s12889-026-29569-2

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29569-2

Keywords: COVID-19, China, zero-COVID policy, crisis communication, government reputation, public trust, health compliance, leadership, transparency, randomized experiment, Crisis, communication