--- Mirko <[email protected]> wrote:
> is there any way to compute the optimal number of classes in a latent
> class analysis using Stata? The only way I am aware of doing latent
> class analysis in by using -gllamm-, but I did not find a way of
> choosing the number of classes other than the comparison of
> likelihood ratio (which is not the best way).
Mirko:
If you have the likelihood you could compare BICs: -2*log-likelihood +
npar*log(nobs). npar is the number of parameters in your model and nobs
is the number of observations. However, never rely on just one
statistic, also have a look at residuals, how much difference changes
in the number of classes make (i.e. how robust the results are), and
last but definately not least just plain common sense and your
substantive knowledge of the situation and the data you are studying.
HTH,
Maarten
-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
___________________________________________________________
Try the all-new Yahoo! Mail. "The New Version is radically easier to use" � The Wall Street Journal
http://uk.docs.yahoo.com/nowyoucan.html
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
