We have been trying to conduct a growth mixture model to establish trajectories of maternal depressive symptoms. We have managed to set up a model in MPlus and it runs, but we have a few theoretical questions that we would really appreciate any help with, since we are relatively new to this type of modelling.

As a background, we have maternal symptom scores measured antenatally, at 2,6 and 12 months, as well as 13 and 17 years of child age. We have two questions relating to this:

  1. does it make sense to model time continuously/ can we use all the time-points, given that there is such a large gap between the first four and last two? We couldn’t find any examples of studies with a comparable data structure, so would be very grateful for any advice.

  2. if we understood correctly, measurement invariance is one of the assumptions underlying these types of models. In our study, early depressive symptoms were assessed via the Edinburgh Postnatal Depression Scale, and for the later two years via the PHQ. These questionnaires have similar scale ranges, but are different in terms of item contents. Is it possible to account for this in some way to still be able to analyse the data?

Any help would be greatly appreciated!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.