I would like to ask you a question about the “cluster” command embedded in Mplus, the command mentioned on pages 456-457 of your book: Mplus user’s guide version 5.
My data are longitudinal (variables are measured at four occasions: Time 1, Time 2, Time, and Time 4). Moreover, the data file is composed of students that are nested in different classrooms (15 classrooms in total with an average of 20 students per classroom).
To test my hypotheses, I used a cross-lag model where latent constructs x and y are expected to be reciprocally related over time. The problem that I have met was that my participants are not nested in the same groups over time. More specifically, for Time 1 and Time 2, students remain in the same classroom over the school year (thus from Time 1 to Time 2 they are exposed to the same teacher). However, from Time 2 to Time 3, they experienced a transition from 10 grade to 11 grade, thereby the composition of the classroom groups is changed (students are no longer exposed to the same teacher). However, from Time 3 to Time 4, they remain exposed to the same teachers and thus the composition of the classroom groups remains the same (but, of course, different from the one from Time 1 and Time 2).
So, my question is how to use the cluster command when students change from one group to another?