Hello,
I am runnning the following model
and I get the direct/indirect effects using the option
which uses the VCE matrix from the previous estimation. This matrix size is e(V) [11,11]
The SE of my second equation however need to be corrected so I need to create a new VCE matrix containing the corrected SE of the second equation.
Since the sem command essentially estimates two OLS regressions, I have proceed as follows:
Step 1 - Estimate VCE from equation 1
The size of vceq1 is [5,5]
Step 2 - Show VCE from equation 2 using corrected SE previously calculated.
The size of vceq2 is [6,6]
What I need to do is to create a new VCE matrix, newvce, that contains the information of vceq1 and vceq2 and with a size [11,11] so that I can run the following command
Could anyone tell me how to create the new matrix newvce?
Thank you very much for your help.
Rubén
I am runnning the following model
Code:
sem(M<-T C1 C2 C3)(Y<- T M C1 C2 C3)
Code:
estat teffects
The SE of my second equation however need to be corrected so I need to create a new VCE matrix containing the corrected SE of the second equation.
Since the sem command essentially estimates two OLS regressions, I have proceed as follows:
Step 1 - Estimate VCE from equation 1
Code:
reg M T C1 C2 C3 matrix vceq1 = e(V)
Step 2 - Show VCE from equation 2 using corrected SE previously calculated.
Code:
matrix list vceq2
What I need to do is to create a new VCE matrix, newvce, that contains the information of vceq1 and vceq2 and with a size [11,11] so that I can run the following command
Code:
sem(M<-T C1 C2 C3)(Y<- T M C1 C2 C3) matrix b=e(b) ereturn post b newvce estat teffects
Thank you very much for your help.
Rubén