Multiple Group Analysis PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
 Dylan John posted on Tuesday, July 25, 2017 - 11:38 am
I am trying to wrap my head on how to conduct a test of measurement invariance in MPlus. My data consists of a latent profile analysis at three timepoints. I am aware that I need to create a longfile dataset where a grouping variable specifies each timepoint. Do I also need to specify the latent profile groupings for each timepoint or is measurement invariance calculated across the whole dataset?
 Dylan John posted on Tuesday, July 25, 2017 - 11:53 am
Here are a few more details about the type of data I have:

I have a single dataset, in a longfile format. There is a variable that specifies ID, observation timepoint (t1, t2, or t3), latent group they are most likely to belong to (within their respective timepoint), as well as the 4 indicator variables for each timepoint (e.g. t1x1, t1x2, t1x3, t1x4...t3x4).
 Bengt O. Muthen posted on Tuesday, July 25, 2017 - 6:09 pm
You should analyze this in wide format - then it is easy to check measurement invariance across time. If your data is in long form you can use our DATALONGTOWIDE command.
 Dylan John posted on Tuesday, July 25, 2017 - 6:37 pm
Ok so you are saying that I should have cases with multiple observations (across timepoints 1-3), on the same line in the dataset?
 Linda K. Muthen posted on Wednesday, July 26, 2017 - 5:53 am
Yes. Each time point is a separate variable. So if y is measured four times, you would have four variables in the data set.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message