TABLE 2.
Measurement invariance models comparing the Italian and Belgian sample.
Model | Compared model | χ2 | df | CFI | ΔCFI | RMSEA | ΔRMSEA | SRMR | BIC | ΔBIC | |
0 | Configural inv. | 235.02 | 160 | 0.956 | 0.044 | 0.054 | 20,456 | ||||
1 | Weak factorial inv. a | 0 | 248.31 | 170 | 0.954 | −0.002 | 0.044 | 0.000 | 0.062 | 20,413 | −43 |
2 | Strong factorial inv. b | 1 | 295.53 | 180 | 0.932 | −0.022 | 0.052 | + 0.008 | 0.070 | 20,405 | −8 |
2.1 | Partial strong factorial inv. c | 1 | 277.97 | 179 | 0.941 | −0.013 | 0.048 | + 0.004 | 0.067 | 20,391 | −22 |
3 | Strict factorial inv. d | 2.1 | 321.96 | 194 | 0.924 | −0.017 | 0.053 | + 0.005 | 0.083 | 20,355 | −36 |
3.1 | Partial strict factorial inv. e | 2.1 | 313.55 | 193 | 0.929 | −0.012 | 0.051 | + 0.003 | 0.079 | 20,351 | −40 |
3.2 | Partial strict factorial inv. f | 2.1 | 304.13 | 192 | 0.934 | −0.007 | 0.050 | + 0.002 | 0.078 | 20,346 | −45 |
4 | Factor (co)var. inv. g | 3.2 | 345.97 | 207 | 0.918 | −0.016 | 0.053 | + 0.003 | 0.103 | 20,309 | −37 |
4.1 | Partial factor (co)var. inv. h | 3.2 | 333.58 | 206 | 0.924 | −0.010 | 0.051 | + 0.001 | 0.099 | 20,299 | −47 |
5 | Factor means inv. i | 4.1 | 385.40 | 211 | 0.897 | −0.027 | 0.059 | + 0.008 | 0.113 | 20,329 | + 30 |
5.1 | Partial factor means inv. j | 4.1 | 342.47 | 210 | 0.922 | −0.002 | 0.052 | + 0.001 | 0.103 | 20,284 | −15 |
CFI, Comparative fit index; RMSEA, Root mean square error of approximation; SRMR, Standardized root mean square residual; BIC, Bayesian information criterion; inv., invariance; (co)var., (co)variance.
a Model with constrained factor loadings across countries. b Model with constrained intercepts and free latent means on top of the constraints of Model 1. c Model as Model 2, with the intercept of item 9 released, so it could vary across countries. d Model with constrained residual variances on top of the constraints of Model 2.1. e Model as Model 3, with the residual variance of item 7 released.f Model as Model 3.1, with the residual variance of item 2 released. g Model with constrained factor (co)variances on top of the constraints of Model 3.2. h Model as Model 4, with the factor variance of Non-Intervention released. i Model with constrained factor means on top of the constraints of Model 4.1. j Model as Model 5, with the factor mean of Group Discussion released.