Skip to main content
. Author manuscript; available in PMC: 2018 Sep 1.
Published in final edited form as: Psychol Med. 2017 Dec 19;48(12):2073–2084. doi: 10.1017/S0033291717003610

Table 4.

analyses of measurement invariance across five samples of the higher-order structural model of common mental disorders.

Model Model parameters Fit indices
CFI RMSEA (90%CI) ΔCFI
1. Configural invariance -1st and 2nd order factor loadings, item-thresholds and factor correlations unconstrained; 0.978 0.019 (0.016–0.022) -
Fixed parameters:

  • Factor means and 1st order factor intercepts=0 across countries for model identification.

  • Factor variances=1 across countries.

  • Item residual variances =1 across countries.
2. Partial invariance -1st and 2nd order factor loadings constrained;
-Item-thresholds and correlations unconstrained.		 0.976 0.018 (0.015–0.021) 0.002
Fixed parameters:

  • Factor means and 1st order factor intercepts=0 across countries for model identification.

  • Factor variances=1 across countries.

  • Item residual variances =1 across countries.
3. Scalar invariance 1st and 2nd order factor loadings and item-thresholds constrained.
-Factor means, factor correlations unconstrained.		 0.962 0.021 (0.019–0.024) 0.016
Fixed parameters:

  • 1st order factor intercepts=0 across countries for model identification.

  • Factor variances=1 across countries.

  • Item residual variances =1 across countries.

CFI=comparative fit index; RMSEA=root mean squared error of approximation.

All models testing invariance across 10 samples.

All models estimated with the theta parameterization in Mplus 7 (Muthén & Muthén, 1998–2012).