Good morning!
In short, we are running a multigroup structural equation model using sem command and group(). We have three groups for this problem. We initially run ginvariant(none) command to check on which parameters needed to be relaxed. When we do so, we run the estat ginvariant command to gather both Wald and score test. According to the tscore, one parameter should be relaxed, so we do so, making the three groups different. This results in the Wald test p value being marginally significant. Subsequently, we constrained two groups to be the same and one to be different, which in theory would left us with 1 df for the Wald test. However, when we ran again the estat ginvariant command, the parameter we changed it is not reported in the table anymore. Does anyone knows why this might be?
Thanks for any insights!
LLLR
In short, we are running a multigroup structural equation model using sem command and group(). We have three groups for this problem. We initially run ginvariant(none) command to check on which parameters needed to be relaxed. When we do so, we run the estat ginvariant command to gather both Wald and score test. According to the tscore, one parameter should be relaxed, so we do so, making the three groups different. This results in the Wald test p value being marginally significant. Subsequently, we constrained two groups to be the same and one to be different, which in theory would left us with 1 df for the Wald test. However, when we ran again the estat ginvariant command, the parameter we changed it is not reported in the table anymore. Does anyone knows why this might be?
Thanks for any insights!
LLLR
