Help with boostrap code for indirect effects using gsem

[gi peters]

Hello,
I am trying to obtain coefficients, bootstrap standard errors and percentile CI for an indirect effect using gsem. I saw this code online

 
 capture program drop boot1  
 program boot1, rclass  
 gsem (Y<- X1 i.X2 X3 Z P[province] C[province>district_no]) ///  
 (Z<-X1 i.X2 X3 P[province] C[province>district_no]), nocapslatent latent (P C)  
 return scalar X1_indirect=(_b[Z:X1]*_b[Y:Z])  
 return scalar X21_indirect=(_b[Z:1.X2]*_b[Y:Z])  
 return scalar X22_indirect=(_b[Z:2.X2]*_b[Y:Z])  
 return scalar X3_indirect=(_b[Z:X3]*_b[Y:Z])  
 end  
 bootstrap r(X1_indirect) r(X21_indirect) r(X22_indirect) r(X3_indirect), seed(12345) reps(200): boot1 estat boot, bc percentile

This works but it instead of just showing the outputs for X1_indirect, X21_indirect, and X3_indirect which it does now, I also want to see the output for these portions of the code

 
 gsem (Y<- X1 i.X2 X3 Z P[province] C[province>district_no]) ///  
 (Z<-X1 i.X2 X3 P[province] C[province>district_no]), nocapslatent latent (P C)

Could someone help me with how to edit the code in a way that it would show me the outputs for these section too?

Thanks in advance!

[Clyde Schechter]

-bootstrap- has a -noisily- option that will show you all the output from each rep.

But what are you going to do with the output from 200 replications of -gsem-? It will be voluminous, and I don't see how you will be able to read and understand it or do anything useful with it.

[gi peters]

Thanks Clyde Schechter but I dont really want to see the output for the 200 replications needed to generate the indirect effects.

I just wanted to see what the outputs for this section is (in addition to the bootstrapped indirect effects)

  gsem (Y<- X1 i.X2 X3 Z P[province] C[province>district_no]) ///    (Z<-X1 i.X2 X3 P[province] C[province>district_no]), nocapslatent latent (P C)

Would that be possible?

I know how to do it without bootstrapping but Im required to use bootstrapping for my dataset so Im having a hard time figuring out the code.
Without bootstrapping, my codes will simply be

gsem (Y<-X1 M controls)(M<X1 controls)
*to get the indirect effect, I will do
nlcom _b[Y:M]*_b[M:X1]
*the code is not directly the same as above but the logic remains

[Clyde Schechter]

Just run that code separately, before or after the bootstrap. I don't think there is any way to get -bootstrap- to show the output of the underlying analysis on the original sample.

If you didn't have random effects in the model, you could switch from -gsem- to -sem-. Then -sem- allows you to specify -vce(bootstrap)- instead of directly using the -bootstrap- command. The output for that would show you the coefficients from the underlying analysis, but the standard errors and its sequel statistics would be calculated by bootstrapping. Unfortunately, -gsem- does not support -vce(bootstrap)-, and -sem- doesn't support group level latent variables.

[gi peters]

Thank you Clyde Schechter . That's my initial thought as well..