Multiple group, multiple indicator gr...
Message/Author
 Christopher Brydges posted on Tuesday, April 10, 2018 - 2:13 pm
I apologise in advance if I misuse any terms, I'm new to Mplus.

My study has two pre-defined groups (control vs. treatment) who were tested pre- and post-treatment on five tasks each time that should load onto the same latent variable (all continuous variables), with six covariates. I want to see if the treatment group had improved performance. Based on my reading, I've combined syntax from examples 6.14 (Multiple indicator linear growth model for continuous outcomes) and 8.8 (GMM with known classes(multiple group analysis)). Due to message size limits, the syntax is in the next post on this thread.

The model runs, but at this point I'm not sure what to do next - at some point, I'm guessing I will constrain some estimates of i to be equal between groups and then determine if estimates of s differ between groups or not to examine the change between the groups.

I have the following questions:
1. Is this syntax doing what I want it to?
2a. If it is, what is the next step? Possibly some kind of invariance testing?
2b. If it isn't, could I be pointed in the direction of the analyses/examples/syntax that are suitable for this.

Thank you very much in advance.
 Christopher Brydges posted on Tuesday, April 10, 2018 - 2:14 pm
Mplus VERSION 8
INPUT INSTRUCTIONS

TITLE: Multiple indicator linear growth model with latent variables and groups
DATA: FILE IS dataset.dat;
VARIABLE: NAMES ARE g x1-x6 y11 y12 y21 y22
y31 y32 y41 y42 y51 y52;
USEVARIABLES ARE x1-x6 y11 y12 y21 y22
y31 y32 y41 y42 y51 y52;
CLASSES = cg (2) c (2);
KNOWNCLASS = cg (g = 0 g = 1);
MISSING = all(999);
ANALYSIS: TYPE = MIXTURE;
MODEL: %OVERALL%
f1 BY y11 y21 y31 y41 y51 (1-5);
f2 BY y12 y22 y32 y42 y52 (1-5);
[y11 y21 y31 y41 y51] (6);
[y12 y22 y32 y42 y52] (7);
i s | f1@0 f2@1;
i s ON x1 x2 x3 x4 x5 x6;
c ON cg x1 x2 x3 x4 x5 x6;
%cg#1.c#1%
[i*2 s*1];
%cg#1.c#2%
[i*0 s*0];
%cg#2.c#1%
[i*3 s*1.5];
%cg#2.c#2%
[i*1 s*.5];
OUTPUT: TECH1 TECH8;
 Bengt O. Muthen posted on Tuesday, April 10, 2018 - 3:38 pm
With only 2 time points, there is no point in doing growth modeling. Simply do a longitudinal CFA where you have measurement invariance over time and where you study change by fixing the factor mean at zero for the first time point and estimate it at the second.

We ask that postings be limited to one window. Longer inquiries should be sent to Support along with license number.
 Christopher Brydges posted on Wednesday, April 11, 2018 - 8:31 am