The Fear of COVID-19 Scale: A meta-Analytic structural equation modeling approach

Abstract

The widespread administration and multiple validations of the Fear of Covid-19 Scale (FCV-19S) in different languages have highlighted the controversy over its underlying structure and the resulting reliability index. In the present study, a meta-analysis based on structural equation modeling (MASEM) was conducted to assess the internal structure of the 7-item, 5-point Likert-type FCV-19S version, estimate an overall reliability index from the underlying model that best reflected the internal structure (one τ-equivalent factor, one congeneric factor, or two-factor models), and perform moderator analyses for the model-implied inter item correlations and estimated factor loadings. A Pearson inter-item correlation matrix was obtained for 48 independent studies, from which a pooled matrix was calculated following a random-effects multivariate meta-analysis. The results from the one-stage MASEM analysis showed that the two-factor model properly fitted the pooled matrix, while the τ-equivalent and congeneric one-factor models did not. Even though, the use of a bifactor model exhibited the predominance of the general factor over the domain-specific ones. High omega coefficients were obtained for the entire scale (.91) and the psychological (.83) and physiological (.83) symptoms subscales. Moderator analyses evidenced an increase in the estimated factor loadings, as well as in the reliability of the FCV-19S, when the standard deviation of the total scores increased and when the FCV-19S was administered to specific (vs. general) populations. The FCV-19S can be therefore considered as a highly related two-factor scale whose reliability makes it suitable for applied and research purposes.