I want to estimate a growth model for scores on a mental health questionnaire administered at five time points (T1-T5). I want to estimate the influence of a covariate measured at T2 only on questionnaire scores at T2-T5, but this covariate will have no impact on questionnaire scores at T1.
Can you perhaps point me to an example in the Mplus user guide that might allow me to model this?
Just let the T2 covariate correlate with the growth factors. Then you can let it predict the T2-T5 outcomes directly or indirectly via the slope growth factor (I assume that the time score is zero at T1).