R McDowell posted on Tuesday, November 07, 2017 - 8:50 am
Hello. This is just a point of clarification on missing/incomplete data in bivariate LGCM. Supposing outcome one is measured for person 1 at a number of time-points but outcome two is not applicable/relevant to this person. Vice versa for person 2. For person 3, both outcomes are measured, although the time points at which the data is available may vary between the two outcomes for this person. When the LGCM model is run, does the FIML estimation method include all available data (ie all three persons will be included), or does it only include persons who have at least some data for both outcomes? Thank you for this clarification.