|
Fit indexes
|
χ2
|
df
|
p
|
χ2/df
|
Δχ2
|
Δdf
|
p’
|
CFI
|
GFI
|
AGFI
|
AIC
|
RMSEA
|
|---|
|
Configural invariance model
|
244.051
|
102
|
.000
|
2.393
|
-
|
-
|
-
|
.911
|
.929
|
.886
|
370.051
|
.048
|
|
Weak invariance model
|
265.043
|
118
|
.000
|
2.246
|
20.992
|
16
|
.179
|
.908
|
.923
|
.893
|
359.043
|
.045
|
|
Strong invariance model
|
301.086
|
120
|
.000
|
2.509
|
57.035
|
18
|
.000
|
.887
|
.912
|
.878
|
391.086
|
.049
|
|
Complete invariance model
|
400.484
|
140
|
.000
|
2.861
|
156.433
|
38
|
.000
|
.837
|
.888
|
.868
|
450.484
|
.055
|
- Equivalences for AMI-Met were tested in sequence by placing constraints on the parameters (i.e., factor loadings, factor variances/covariance, and error variances). Configural invariance is the simplest form without any constraints on the parameters. Weak invariance constrains the factor loadings; strong invariance constrains the factor loadings and factor variances/covariance; and complete invariance constrains the factor loadings, factor variances/covariance, and error. The difference in the χ2 values (Δχ2) and the degrees of freedom (Δdf) were calculated as follows: Δχ
2 = χ
2
nested model − χ
2
configural invariance model, and Δdf = dfnested model − dfconfigural invariance model. p’ indicates the significance distributed from Δχ2 and Δdf [24]
