Objective   To evaluate the reliability and validity of a self-designed scale on perceived risk of coronavirus disease 2019 (COVID-19) epidemic among Chinese citizens with an online empirical study.

  Methods  Theoretical framework and items of the COVID-19 Risk Perception Scale were determined using literature researches and expert consultations. The designed scale was administered via an online platform among the public in 31 provincial level administrative divisions (PLADs) across China. Statistic analyses on valid information collected from a total of 19 132 respondents were performed to evaluate the scale's content validity and construct validity; Cronbach's α coefficient and split-half reliability coefficient were adopted in reliability assessment and confirmatory factor analysis was used to evaluate the fitness of the constructed model. Regional disparity in COVID-19 epidemic risk perception of the public was explored with one-way analysis of variance (ANOVA) using SPSS 24.0.

  Results   The constructed scale was consisted of nine items for three dimensions of susceptibility, severity, and controllability of the epidemic. For the whole scale, the Cronbach′s α coefficient is 0.824, and the split-half reliability coefficient is 0.731. The correlation coefficients between each dimension and the whole scale ranged 0.720 – 0.841. Confirmatory factor analysis showed that the fitting indicators of the constructed model were all within the standard range, with a goodness-of-fit index (GFI) of 0.982, a comparative fit index (CFI) of 0.972, and a root mean square error of approximation (RMSEA) of 0.062, indicating good reliability and validity of the scale developed. A "ripple effect" in the epidemic risk perception was observed among respondents in different PLADs with diverse prevalence of COVID-19 based on the survey with the scale.

  Conclusion  The constructed scale is of good reliability and validity and could be used in the measurement on COVID-19 epidemic risk perception among Chinese citizens.