Fig. 1

Model fit of the HLS-EU-Q47 after applying various analysis approaches. Figure 1 shows the overall fit statistics for the one-dimensional approach (all subscale correlations fixed to 1), the consecutive approach (treating the three health domains as orthogonal or uncorrelated) and the two-, three- and 12-dimensional approaches (treating the theoretical subscales as correlated). A: access, B: understand, C: appraise, D: apply (cognitive domains). HC: health care, DP: disease prevention, HP: health promotion (health domains). Δ: change in parameter, AIC: Akaike’s information criterion, cv: critical value, D: deviance, df: degrees of freedom, ep: number of estimated parameters, LRT: likelihood ratio test. PSR: person separation reliability based on marginal maximum likelihood estimate/Warm’s mean weighted likelihood estimate