Hi,
In the univariate ACE modeling, I constrained the total variance of A, C and E as 1. As suggested by Rob, I put the following lines in the scripts:
unit <- mxMatrix( type="Unit", nrow=1, ncol=1, name="Unit") identifyingConstraint <- mxConstraint(VarP == Unit, name="ic")
I found that the results have been standardized that means the a2 is equils to the a2/(a2+c2+e2). However, the results is different from that if the total variance was not constrained as 1. Furthermore, if I did not constrain the total variance as 1, the scripts always echo red error 5 though the results seems correct. Hence I'm wondering that if to constrain the total variance as 1 is necessary and why? Many thanks~