Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Growth Mixture Modelling commands in gllamm


From   Cheryl Stopford <Cheryl.Stopford@manchester.ac.uk>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: Growth Mixture Modelling commands in gllamm
Date   Thu, 21 Feb 2013 09:47:19 +0000

Hi,

I have four years' worth of data and want to look at variation in 
individual patterns over the four years. Cross-sectional plots of each 
year of data show there's an overall pattern according to the stage of 
disease where peoples' scores increase for the first
 two stages, then decrease, then increase again. The original hypothesis
 when the data were viewed cross-sectionally was that the 'decrease' at 
stage 3 was specific to that stage. However the longitudinal analysis 
might show something different - for example,
 that it's the same individual participants that are scoring at a high /
 low level at each visit.

Scores on the scale I'm using are both positives and negatives. I've 
explored the data using both repeated cross-sectional cluster analyses 
(to look at variation in patterns) and, for longitudinal analysis, 
linear mixed modelling. The pattern I'm looking at
 doesn't appear to be particularly linear, so I'm not sure that linear 
mixed modelling is the best method to use.

Initial exploratory findings seem to show that there are four distinct 
patterns in the data occurring longitudinally (high scorers, low 
scorers, and two intermediate groups that score around the mean). I'd 
really like to know if the same individuals have the
 same pattern longitudinally. I'm pretty convinced that I need to use a 
Growth Mixture Modelling technique to look at the data in a meaningful 
way but I'm really not very sure how to go about it.

I'm very new to stata and have only just recently got to grips with the 
basics for the purposes of linear mixed modelling. I've looked at the 
GLLAMM manual (http://biostats.bepress.com/cgi/view...ext=ucbbiostat)
 but am just finding it really complicated to follow with my rudimentary knowledge...!

I have three main questions that seem to have arisen over the course of my reading:

1) when you do Growth Mixture Modelling in GLLAMM, what format should 
the data be in? Previously I'd restructured it for multilevel modelling 
so that each participant's different visit was recorded on a different 
line, but something I read in the GLLAMM manual
 suggested otherwise for growth mixture modelling - is this right?

2) I know it's a big ask, but can anyone help to modify my xtmixed mixed
 level modelling command into a gllamm growth mixture modelling command?
 Data can be nested into two levels (participant, which looks like it 
accounts for 60% of the variability; and original
 disease subgroup, of which there were 4).

If it helps at all, my original xtmixed command looks like this:
xtmixed score time subgroup || Participant: time, cov(uns) mle var if Data_points>=3

3) If not, can anyone direct me towards a simple explanation of GLLAMM 
procedures (without too much technical background reading!!)

Thanks for your time and help with this! 

Cheryl







*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index