[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Latent class estimation

From   Phil Schumm <[email protected]>
To   [email protected]
Subject   Re: st: Latent class estimation
Date   Wed, 5 Dec 2007 09:35:29 -0600

On Dec 5, 2007, at 4:57 AM, Eva Poen wrote:
I am interested in running a latent class regression model.

On Dec 5, 2007, at 7:01 AM, Dimitris Pavlopoulos wrote:
I have not explored the potential of STATA in latent class models, but I would suggest you to do your analysis in Latent Gold instead. This is a program especially made for such analysis. And yes, you need to use the EM algorithm and in particular its modified version (Baum Weltch algorithm or something like that). This is implemented in Latent Gold.

On Dec 5, 2007, at 7:21 AM, Christian Deindl wrote:
I think the best book for beginners in LCA is still:

McCutcheon, A. L. (1987). Latent Class Analysis, Beverly Hills: Sage Publications.

Apart from latent gold you can also use lem ( faculteiten/fsw/organisatie/departementen/mto/software2.html) in the beginning and I'm also pretty sure that GLLAMM will be useful, if you want to do your analysis with stata.
On Dec 5, 2007, at 7:32 AM, David Airey wrote:
Another program that can be used for just about any latent variable modeling question whether continuous or categorical, even multilevel, is Mplus, now at version 5.

I would like to offer an alternative opinion. RE references that are accessible, Bartholomew's "Latent Variable Models and Factor Analysis" (1987, Charles Griffin & Co.) offers a good theoretical introduction, and Skrondal and Rabe-Hesketh's "Generalized Latent Variable Modeling" (2005, Chapman & Hall) is superb in terms of both its theoretical and applied presentation. The latter should be read in any case without question.

The program -gllamm- (type -ssc install gllamm-) is capable of estimating latent class models in a very flexible manner, and the manual available with the program (as well as the book cited above) offer several worked examples. It can be slow for datasets with a large number of unique covariate patterns, but this has to be evaluated relative to the time required to learn how to use another program and to move your data and results back and forth. It's certainly an excellent place to start and to run some preliminary analyses; you can always then move to another piece of software if speed becomes a problem.

-- Phil

* For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index