|
|
Multilevel data generation |
|
Message/Author |
|
|
Hi, I want to generate multilevel data (2 levels) of a CFA model with categorical outcome variables using the monte carlo option. I understood that the within error variance is not a free parameter in this case and that it is automatically fixed to 1 in combination with a probit link. Is that correct? Am I correct assuming then, that when estimating this same CFA model with WLSMV, the underlying implicit parameterization is the Theta parameterization (as it does not seem to be possible to choose between theta and delta in this case)? Thanks in advance! |
|
|
Yes and yes. |
|
Back to top |
|
|