Multigroup comparison
Message/Author
 Sofia Diamantopoulou posted on Thursday, May 20, 2010 - 2:14 am
Hi,

I have conducted a latent class growth analysis and the best fitting model indicates the existence of three separate trajectories differing in i, s, and q growth factor. Now I would like to test whether this model fits equally well for boys and girls. In other words I need to test whether a three trajectory class model fits equally well for boys and girls and if so, whether the intercepts and growth factors differ by gender. Could you give me a tip on how to conduct the multigroup analysis?
 Jon Heron posted on Thursday, May 20, 2010 - 2:49 am
Hi Sofia,

you want to use the KNOWNCLASS option. Check out example 8.8 (p216 in the v6 guide)

best wishes, Jon
 Amy Tobler posted on Thursday, September 02, 2010 - 7:38 am
We have a situtation where we wish to compare the growth rates between two groups (control and experimental) and we had thought to run this as a multigroup growth model. We also have multilevel data (students nested within school units). We tried the following code:

VARIABLE: NAMES ARE ALCYEAR1 ALCYEAR2 ALCYEAR3 ALCYEAR4 trt nunit fam lunch;
CATEGORICAL = ALCYEAR1 ALCYEAR2 ALCYEAR3 ALCYEAR4;
WITHIN=fam lunch;
CLUSTER= NUNIT;
CLASSES=c1(2);
KNOWNCLASS= c1 (trt=0 trt=1);
MISSING are .;
ANALYSIS: TYPE= TWOLEVEL MIXTURE ;
MODEL:
%WITHIN%
%OVERALL%
iw sw | ALCYEAR1@0 ALCYEAR2@6 ALCYEAR3@18 ALCYEAR4@30;
iw sw ON fam lunch;
%BETWEEN%
%OVERALL%
ib sb | ALCYEAR1@0 ALCYEAR2@6 ALCYEAR3@18 ALCYEAR4@30;
OUTPUT: TECH1;

And recieved the following error:
*** FATAL ERROR THIS MODEL CAN BE DONE ONLY WITH MONTECARLO INTEGRATION.

Any suggestions you can give on either fixing our code or an easier way to answer our question would be appreciated.
 Linda K. Muthen posted on Thursday, September 02, 2010 - 7:43 am
Add INTEGRATION=MONTECARLO; to the ANALYSIS command.