*2.3. Data Analysis*

Pearson's correlations were calculated to examine the relationships among the variables. Then, structural equation modelling (SEM) was used to estimate the statistical models (Amos 21 SPSS, an IBM Company: Chicago, IL, USA) [47]. In accordance with the rules of SEM, we first validated the measurement model by confirmatory factor analyses (CFA) and then we tested the structural relationship between measured variables and latent constructs. The goodness of fit of the model was evaluated by applying different indices: χ<sup>2</sup> statistic, the standardized root-mean-square residual (SRMR), root-mean-square error of approximation (RMSEA), and three modification indices—goodness of fit index (GFI), Tucker Lewis index (TLI), and comparative fit index (CFI). We were removing non-significant parameters and non-significant paths of the original model one at a time in relation to the modification indices to improve the model fit (Table S1, Supplementary Material) [48]. In addition, the modifications were to improve the model fit as long as they were corroborated by the theory. Bootstrapping was used to estimate direct and indirect effects (samples = 5000; 95% bias-corrected confidence intervals; standardized coefficients were presented). This approach enabled us to examine the relationship between risk perception of COVID-19 and psychological well-being as well as the relationship between meaning-based resources and psychological well-being.
