|
|
 |
| GGMM with synthetic cohort data |
 
|
| Message/Author |
|
|
|
|
Greetings all,
New to the forum, have owned MPlus for a while, waiting for my chance to put it to work.
I'm working on a study of adolescents (age 13-19) with 3 (annual) data points for each of 2000 participants and a continuous outcome measure of problem video gaming.
To get a single curve for our outcome we are using Mirowsky & Kim's (ref below) age-vector techniques for identifying any potential cohort effects that might compromise the findings. All this can be carried out in Mplus.
What I wanted to ask, however, is if multiple trajectories in outcome means (including our hypothesized 'escalating,' 'abating,' and 'persistent' trajectory types) can be identified with GGMM in MPlus using such accelerated (also: synthetic or virtual) longitudinal cohort data.
Also, if there are good FAQs on power analysis for GGMM archived here or elsewhere, I would love a link.
With thanks,
Luther Elliott, Ph.D.
National Development and Research Institutes
71 W. 23rd St. New York, NY 10010
REFERENCE
Graphing age trajectories: Vector graphs, synthetic and virtual cohort projections, and cross-sectional profiles of depression.
Mirowsky, John; Kim, Jinyoung
Sociological Methods & Research, Vol 35(4), May 2007, 497-541 |
|
|
|
|
Sounds like you are interested in a multiple-group approach to multiple cohort analysis in line with UG ex6.18. This can be generalized to GMM where you will represent the multiple groups by using the Knownclass option, so you have a combination of a knownclass latent class variable and an unknown (regular) latent class variable.
UG ex 12.3 gives a start for the input for a power study with GMM. I can't think of a writeup on this off hand. |
|
|
|
|
This is a great start, and the cites from the user guide much appreciated. Big thanks!
Luther |
|
|
| Back to top |
|
|
