Evaluating measurement (non)-invariance of the Secure Flourishing Index: A Bayesian moderated factor analysis approach

Background. Measurement invariance is crucial for cross-cultural research, yet identifying specific sources of non-invariance in multidimensional constructs remains methodologically challenging. We applied Bayesian moderated factor analysis to identify sources of measurement non-invariance in the 12-item Secure Flourishing Index across 23 countries using Wave 2 Global Flourishing Study data (N=124,993).Methods. We tested for non-invariance across countries and person characteristics (age, gender) using spike-and-slab priors. Critically, we simultaneously modeled measurement non-invariance and impact on the full latent structure, including factor correlations—a first for multidimensional (6-factor) models.Results. The six-factor structure was largely invariant across groups. Age moderated item intercepts (5/12 items), residual variances (6/12 items), and factor correlations. At least one dummy-coded respondent country identifier moderated at least one parameter for each item. Factor correlations were also moderated. Conclusion. Results support partial invariance of the SFI and demonstrate how moderated factor analysis can reveal measurement heterogeneity unavailable from traditional approaches.