Mplus user guide example 11.4
Message/Author
 Michael Lawton posted on Friday, September 15, 2017 - 6:23 am
I am relatively new to Mplus and pattern-mixture modeling (PMM) and have a question about the PMM example in the Mplus user guide, that is example 11.4.

In a non Structural Equation Modelling framework when carrying out a PMM one would need to carry-out post estimation manipulation of the parameter estimates to get the average intercept and slope (see following document http://www.ssicentral.com/supermix/Documentation/l2Sls.pdf).

Is this not necessary within Mplus? For instance in example 11.4 (see output file on mplus website) is the average intercept accounting for dropout 1.903 and the average slope -1.123? So I do not need to use the estimates of I on d1-d5 post estimation to calculate the average intercept?

 Bengt O. Muthen posted on Friday, September 15, 2017 - 6:08 pm
The intercept of e.g. i is not the mean of i. If you want the mean i and s you have to take into account the effects of the d variables. You can do this using the Model Constraint command with parameter labels given in the Model command.
 Michael Lawton posted on Monday, September 18, 2017 - 2:26 am
Many thanks for your quick response. Is this what I would need to do to get the mean of I?

MODEL: i s | y0@0 y1@1 y2@2 y3@3 y4@4 y5@5;
i ON d1-d5 (p1-p5);
s ON d3-d5;
s ON d1 (1);
s ON d2 (1);
[i] (a);
[d1-d5] (m1-m5);
MODEL CONSTRAINT:
NEW(int);
int = a + m1*p1 + m2*p2 + m3*p3 + m4*p4 + m5*p5;
 Bengt O. Muthen posted on Tuesday, September 19, 2017 - 3:23 pm
This looks right.