LCA with many covariates PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 John Mathew posted on Wednesday, December 12, 2007 - 10:01 am
Prof. Muthen

I am very new to MPlus as we recently bought it.

I have the following model to estimate

Y1 = f(eta1)
Y2 = f(eta3)
Y3 = f(eta2)
eta1 = g(X, eta2, eta3),

where the vector Y1, Y2 and Y2 are the sets of indicator variables for
the latent variable eta1, eta2 and eta3 respectively, and X is a
vector of strictly exogenous variables.

Our model suggests eta1 is likely to come from a multimodal
distribution. I have to do a latent class analysis on eta1.

My questions to you are

Q1. could I fit a "latent class analysis with covariate" type model
when we have continuous latent variables in the covariate vector.

Q2. Could you kindly suggest me some references where they dealt
situations like mine

Thank you,
John
 Bengt O. Muthen posted on Wednesday, December 12, 2007 - 10:17 am
Q1. Yes - especially class-varying means of eta1 would orchestrate the multimodality.

Q2. This has been looked at in the marketing lit, e.g.

Jedidi, K., Jagpal. H.S. & DeSarbo, W.S. (1997). Finite-mixture structural equation models for response-based segmentation and unobserved heterogeneity. Marketing Science, 16, 39-59.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: