Hi, I have unequally spaced longitudinal data. I would like to use the phantom variable approach to introduce additional latent variables to fill in the gaps in the panel design despite the lack of manifest variables. Can I do this in Mplus by applying approriate parameter constraints and if so could you point me in the direction of an example?
I don't see why you want to take that approach to unequally spaced long'l data. I am sure you can set it up in Mplus and perhaps you can go by an approach such as McArdle's latent difference modeling. I don't have a setup handy - haven't tried it myself. Make sure the model is identified.