Measurement Invariance in LCA PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 KB posted on Tuesday, June 13, 2017 - 12:28 am
Hello Dr. Muthen,

I have twelve binary latent class indicators to identifying my subgroups. I observed 3-class model fits the data well. I want to use gender "gend" as a measurement invariance to examine differences in the latent classes. Below is my model to accomplish the task.

I have few questions:
1. Is the model correct to accomplish the task?

2. Which is correct when using a restricted model to compare model indices where I constrained the parameters across groups (i.e., I will not include c ON cg in my model)?

classes = cg(2) c(3);
vs.
classes = c(3) cg(2);


Thank you and hope to hear from you!
KB

...............................
!Model 1: (Unrestricted model: All parameters allowed to vary across group)

...
variable: names bk a b c d e f g h i j k l gend;
usevariables = a-l;
categorical = a-l;
knownclass = cg(gend = 0 gend =1);
classes = cg(2) c(3);
missing = all(-999);
cluster = bk;
analysis: type = mixture complex;
processor = 10;
starts = 1000 50;
model: %overall%
c ON cg;

.........................................
 Bengt O. Muthen posted on Tuesday, June 13, 2017 - 6:26 pm
Your input is correct. The order should be cg(2) c(2) when c is regressed on cg.
 KB posted on Tuesday, June 13, 2017 - 7:15 pm
Thank you, Dr. Muthen for your response. Does it mean either cg(2)c(2) or c(2)cg(2) can be used when c ON cg is not in the model?
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: