Multivariate Outcomes Model - Bayesia... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Lindsay Demers posted on Monday, April 04, 2011 - 7:23 am
Hi,
I am currently analyzing data using a multivariate outcomes application of HLM (e.g., see Brennan, Kim, Wenz-Gross et al. 1997). However, I would like to apply Bayesian estimation to this model. As I've been running it, the within-subject model looks like this:
yij = b1(outcome 1) + b2(outcome 2) + b3(outcome 3) + r
Where outcomes 1-3 are 0/1 dummy codes and b1-b3 are treated as latent variables. To do this I've been using a known reliability for each of these outcomes calculating the measurement error myself (1-reliability) * observed variance. Also, note that there is no intercept because the model is saturated.
The between-subject model includes one continuous covariate. So, in HLM terms it's:
B1 = G10 + G11(Covariate) + u1j
B2 = G20 + G21(Covariate) + u2j
B3 = G30 + G31(Covariate) + u3j

Although I intend to take the Bayesian short course at UConn, I would like to analyze these data ASAP with Bayesian estimation. Could you please advise me as to what commands I should use in MPlus, given that this is a "special" application of a MLM?
Best,
Lindsay
 Linda K. Muthen posted on Monday, April 04, 2011 - 8:57 am
You would use in the ANALYSIS command:

TYPE=TWOLEVEL;
ESTIMATOR=BAYES;
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: