EFried posted on Friday, August 16, 2013 - 8:36 am
Dear Dr Muthens,
I'd like to investigate whether 9 ordinal items (0,1,2,3) of a questionnaire react differently to an event that happens between two measurement points (n=1500). I don't have a control group, all subjects experience that severe stressor.
Descriptive statistics show that while all items increase significantly, some do so much more than others.
The question would be how to do this in a SEM in MPLUS. I was thinking about testing whether the 9 item change scores differ in a multivariate way, or test whether the nine slopes differ. However, I would not know how to do this.
Would you be able to recommend a procedure from the MPLUS manual, or literature?
You could consider a factor model if that fits the data. If measurement invariance holds, that would mean that all variables change proportionally. For other suggestions, try posting on a general discussion forum like SEMNET.