Skip to main content
. 2022 Jul 14;40:158–166. doi: 10.1016/j.apnu.2022.07.016

Table 4.

Multigroup CFAs according to gender, age, and country.

χ2 df p χ2/df RMSEA (CI) CFI IFI SRMR Comparisons ∆RMSEA ∆CFI ∆SRMR
Measurement invariance according to gender
 Configural invariance 1802.89 176 <.001 10.24 0.058 (0.056;0.060) 0.925 0.925 0.038 NA
 Metric invariance 1920.27 204 <.001 9.41 0.055 (0.053;0.058) 0.921 0.921 0.048 Conf. Vs Metric 0.003 0.004 0.010
 Scalar invariance 2392.59 220 <.001 10.87 0.060 (0.058;0.062) 0.926 0.927 0.072 Metric Vs Scalar 0.005 0.006 0.024
 Error variance invariance 1797.35 192 <.001 9.35 0.055 (0.053;0.058) 0.923 0.923 0.040 Scalar Vs Error 0.005 0.003 0.032
Measurement invariance according to age
 Configural invariance 1997.06 176 <.001 11.34 0.061 (0.059;0.063) 0.920 0.921 0.039 NA
 Metric invariance 2154.94 204 <.001 10.56 0.059 (0.056;0.061) 0.915 0.915 0.045 Conf. Vs Metric 0.005 0.006 0.006
 Scalar invariance 2585.25 220 <.001 11.75 0.059 (0.057; 0.062) 0.916 0.916 0.046 Metric Vs Scalar 0.000 0.001 0.001
 Error variance invariance 2220.93 192 <.001 11.56 0.062 (0.059; 0.064) 0.912 0.912 0.043 Scalar Vs Error 0.003 0.004 0.003
Measurement invariance according to country
 Configural invariance 1614.49 440 <.001 3.66 0.060 (0.057; 0.063) 0.920 0.921 0.048 NA
 Metric invariance 1878.41 552 <.001 3.40 0.057 (0.054; 0.060) 0.910 0.910 0.063 Conf. Vs Metric 0.003 0.011 0.015
 Scalar invariance 3631.68 616 <.001 5.89 0.062 (0.059; 0.064) 0.918 0.918 0.108 Metric Vs Scalar 0.005 0.005 0.045
 Error variance invariance 1791.64 504 <.001 3.55 0.059 (0.056; 0.062) 0.916 0.917 0.052 Scalar Vs Error 0.003 0.002 0.056

Note. CFA = confirmatory factor analysis; χ2 = Satorra-Bentler chi-square; df = degrees of freedom; p = general model significance; χ2/df = normed chi-square; RMSEA = root mean square error of approximation; CFI = comparative fit index; IFI = incremental fit index; SRMR = standardized root mean square residual; ∆ RMSEA = change in RMSEA compared with the previous model (expressed in absolute values); ∆ CFI = change in CFI compared with the previous model (expressed in absolute values); ∆ SRMR = change in SRMR compared with the previous model (expressed in absolute values).

The table shows the four levels tested of measuring invariance (configural, metric, scalar and error variance) to observe if model fit indices remain similar in participants of different gender, country or age group (adolescents aged 12–17 or young adults aged 18–25). For ease of comparison, the increase in the indexes is also showed.