|
|
Mixture sem models with longitudinal ... |
|
Message/Author |
|
|
I am new to mplus. I have to implement mixture sem models using panel data (with repeated measures). I assume that this is possible. Am I correct? Thanks! |
|
|
Mplus can estimate growth mixture models. |
|
|
Dear Mplus support, I estimated a longitudinal factor analysis where my latent variable SAD is measured with 3 observed variables (lung1, lung2, lung3) at 3 different timepoints (see the code below). I would like to perform a mixture model where for each SAD1, SAD2, SAD3 I would like to link a categorical variable C1 (for SAD1), C2 and C3 (in order to obtain 3 clustering at 3 timepoints based on SAD latent variables). How can I do it in the MODEL statement? Many thanks Marco Pannacci MODEL: %OVERALL% SAD1 by lung11* lung21 lung31@1 (d1-d3); SAD2 by lung12* lung22 lung32@1 (d1-d3); SAD3 by lung13* lung23 lung33@1 (d1-d3); |
|
|
Take a look at UG ex 8.7. Although this refers to a growth model, your factor model is similar enough. The different classes can have different factor means if you add measurement invariance with respect to the intercepts of the factor indicators. |
|
Back to top |
|
|