Time varying covariates PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Joseph Foster posted on Thursday, November 01, 2012 - 4:10 pm

I have an issue, and I am certain that I have to find a work around, but I want to check here first.

I have 4 time points for dependent variable (2 in each year)
My time points for time varying covariates are parallel to the first 2 time points of the dependent variable. However, my third time point is an average of several measurement occasions across that year. So, it's kind of like time point 3.5 or at least that is how I have specified it in autoregressive latent panel models in the past.

My hunch is I can't use these as time varying covariates in MPlus or SAS, and that I should just average covariates across all measurement occasions and include as time invariant.

 Bengt O. Muthen posted on Thursday, November 01, 2012 - 9:26 pm
It is a bit problematic, but turning into a time invariant covariate seems drastic. Perhaps you could let that 3rd, averaged time-varying covariate influence the last 2 time points with the same coefficient as an approximation.
 Abas Mirzaei posted on Wednesday, November 07, 2012 - 5:05 pm
I am struggling with time varying covariates, time scores and time points.
I tried to get my head around it, but I am confused!
I want to use ALT model to track the change in my dependent variable. I know that the change in my dependent variable (repeated measure) is not linear. It goes up and down, since it is affected by too many factors.
I want to track its trend from 2000 to 2007.
I know that its change is not quadratic, cubic, nor exponential.
I was just wondering if you please help me with this. I have spent 8 months self learning Mplus, and now I am not quite sure if I am able to use this software or not!

Once again my aim is to estimate the change in the Repeated Measure over time, and I know that this change is not linear.

Any help is much appreciated.

 Linda K. Muthen posted on Thursday, November 08, 2012 - 9:52 am
Please send an output with TYPE=BASIC and MODEL command along with your license number to support@statmodel.com.
 Abas Mirzaei posted on Friday, November 09, 2012 - 6:29 am
I will!

 Donald A. Lloyd posted on Thursday, December 13, 2012 - 1:27 pm
Hello. I want to run a growth mixture model on BMI for a group of subjects and I’d like to use individual age as the time variable.
1. The first question is what type of analysis should I specify in Mplus? If I specify TYPE=MIXTURE, it doesn’t allow me to use individual varying time.
2. In a growth model using individual time, how to specify the cubic curve?

Thanks for any help!
 Linda K. Muthen posted on Thursday, December 13, 2012 - 2:28 pm
I think AT can be used with MIXTURE. To specify a cubic model, use four growth factor names before AT. This is described in the user's guide under AT.
 Donald A. Lloyd posted on Thursday, December 13, 2012 - 3:20 pm
Many thanks. I'll try that. DL.
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