Negative residual variance PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Daniel Rodriguez posted on Saturday, August 31, 2013 - 5:39 pm
Good evening,
I am running an associated processes model and I would like to covary the residual variances for the trend factors; I have level to trend effects which are fine. However, my residual variance for one of the trend factors is negative, but not significant. I of course can constrain it to equal zero and that works fine; I already tried that. However, I am interested in interpreting the relation between the trend factors. Is it possible to interpret the relation between the trend factors even if one of the variances is negative though not significant? Or is that result nonsense?
 Linda K. Muthen posted on Sunday, September 01, 2013 - 6:58 am
You cannot leave a negative residual variance. That makes the model inadmissible. If you fix it to zero, it cannot covary with anything.
 Daniel Rodriguez posted on Sunday, September 01, 2013 - 3:26 pm
Thanks, I agree. I can't seem to get rid of that negative variance no matter what I do. Fixing it to zero seems to be the only option.
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