Attachment | Size |
---|---|
07_Analysis_Independnet2ACE_141031.R [6] | 12.01 KB |
Hi,
I'm attempting to run a BG script of an independent pathways model on 9 dichotomous (symptom) manifest variables. That is, the model includes biometric factors specific to each manifest variable and common to all manifest variables. I've been able to get this model to run without any problems.
Based on some theoretical and empirical work, I'm also attempting to run a similar model with two correlated A factors; one accounts for the genetic variance in three manifest variables, with the other accounts for the genetic variance in the remaining six variables. This is where I'm running into some problems.
I've specified the two A factors as follows:
Ac1Free <- c(rep(T,6),rep(F,3))
Ac2Free <- c(rep(F,6),rep(T,3))
Ac1Values <- c(rep(1,6),rep(0,3))
Ac2Values <- c(rep(0,6),rep(1,3))
pathAc1 <- mxMatrix( type="Full", nrow=nv, ncol=nf, free=Ac1Free, values=Ac1Values, labels=labFull("ac1",nv,nf), name="ac1" )
pathAc2 <- mxMatrix( type="Full", nrow=nv, ncol=nf, free=Ac2Free, values=Ac2Values, labels=labFull("ac2",nv,nf), name="ac2" )
And I've attempted to specify that their correlated as follows:
pathRac <- mxMatrix (type="Diag", nrow=1, ncol=1, free=TRUE, values=0.5, label="rac1", lbound=0, ubound=1, name="rac")
covAc <- mxAlgebra( expression=ac1 %% rac %% t(ac2) + ac2 %% rac %% t(ac1), name="Ac" )
This model runs, but I'm fairly certain I'm doing something wrong because my degrees of freedom and estimated parameters are unchanged from the first model.
Any input on where I might be going wrong would be greatly appreciated.
Thanks,
Jarrod