Covariances among growth parameters i... PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Dana Garbarski posted on Thursday, September 29, 2011 - 6:01 pm
I'm running a latent growth curve model for a binary repeated measure with 9 data points. This model fits well when modeled as a quadratic growth curve with the variance of the quadratic slope factor=0. However, the intercept and (positive) linear slope growth factors have a negative covariance, where theory tells me that they should be positively related. Is this a possible sign of model misspecification? Any thoughts are most appreciated!
 Linda K. Muthen posted on Thursday, September 29, 2011 - 7:34 pm
You can try centering the time scores at the fifth time point. I believe this is discussed in the Raudenbush and Bryk book.
 Dana Garbarski posted on Friday, September 30, 2011 - 12:24 am
Thank you, this was very helpful and did the trick for the LCGM. However, when I run an autoregressive latent trajectory model with the same set of variables, the covariance between the intercept and slope is negative no matter where I center the time scores. Since centering yielded a positive covariance between the growth parameters in the LCGM, I'm wondering if there is something else that would need to be done with an ALT model that I'm missing? Again, I appreciate any suggestions you may have.
 Bengt O. Muthen posted on Friday, September 30, 2011 - 3:17 am
Not sure about that in the ALT context.
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