|
|
 |
| Cluster randomization and GMM |
 
|
| Message/Author |
|
|
| Sisi Ma posted on Thursday, October 19, 2017 - 3:08 pm
|
|
|
Hi,
I have longitudinal data collected on about 200 students from 30 school programs. Each school program is randomized to administer manipulation A or manipulation B to students that belong to that school program.
I would like to build a GMM to characterized the students response over time. How do I take into account that the students are nested within school programs?
Below are my code without taking into account the nesting within school programs.
==============================
Title: GMM
Data: file is 'file.dat';
Variable: names are Treatment T1 T2 T3 T4;
missing=*;
CLASSES=c(3);
Analysis: type=MIXTURE;
STARTS=200 10;
STITERATIONS=10;
Model: %OVERALL%
i s|T1@0 T2@2 T3@3 T4@4;
c#1 ON Treatment;
c#2 ON Treatment;
Output: sampstat standardized tech1 tech8 tech11;
PLOT: SERIES=T1 T2 T3 T4 (s);
TYPE=PLOT3;
===================================
Thanks in advance for your help! |
|
|
|
|
| See the UG examples 10.8 - 10.10. |
|
|
| Back to top |
|
|
