I attended Topic 7 Multilevel Modeling with latent variables using M-Plus workshop at Johns Hopkins, I have a little question on slide pp141 of your presentation. The categorical variables are u1-u6, they were treated as categorical factor indicators on the WITHIN level. However, on the between level, they were continuous latent variable. May I know why is that so? Appreciate if you could advise. Also where could I find the example data sets used in the presentation?
On the Between level u1-u6 are random intercepts and therefore continuous latent variables varying across Between-level units.
The example is on the Mplus CD for UG ex 9.9 and also on the Mplus web site under Documentation.
Huang Wu posted on Thursday, June 21, 2018 - 8:15 pm
Hi, Dr. Muthen,
I have some questions regarding multilevel mediation analysis with latent variables.
The mediator(M) is in first level and is an latent variable. I want to investigate the x(second level) effect on y(first level) through M(first level). How should I specify the model? I have tried many times, the software told me that the within level factor cannot be used on between level, in this situation, do I need to define an between level factor? Please help, thanks