Latent growth model with seqential-cohort design


New Member

I am trying to run a latent growth model with four time points using an accelerated or seqential cohort design.
One group has data for the first 3 measruement points and the second group provides data for measurement points 2-4.

As I understand running such a model should be possible in AMOS when using a multigroup analysis.

However, when I try to run the model I contantly get an error message that the model cannot be computed because on one of the variables all the values of one group are missing (which is indeed the case as this is the idea behind this design).
I have tried to set all the values (regression weights, errors) related to these variable to zero in their corresponding group (i.e. the values for the last measurement for the first group and for the first measurement for the last group), but that still does not seem to help.

Does anyone have any other ideas?