R help - Aug 2004 - NLME: Holding constant the across group correlational structure of the fixed effects in nlme

Hello all.
I was wondering if there is a way to hold constant the fixed effects correlation
structure across multiple groups?

For example, I have two groups and I fit a three parameter logistic growth curve
where the fixed effects are free to vary across the groups. I'll paste in
the code as a concrete example:
> Result.NLME <- nlme(Score ~ SSlogis(Time, Asym, xmid, scal),
+                            data=Grouped.Data, 
+                            fixed=Asym + xmid + scal ~ Est.LC,
+                            random=Asym + xmid + scal ~ 1,
+                            start=c(105,-5,
4,-5,30,-5))
> Sum.Result.NLME <- summary(Result.NLME) 
> 
> round(Sum.Result.NLME$corFixed, 2)
Asym.(Intercept)  Asym.Est.LC    xmid.(Intercept)    xmid.Est.LC   
scal.(Intercept)     scal.Est.LC
Asym.(Intercept)        1.00          -0.93                0.55                 
-0.51                 0.50                    -0.46
Asym.Est.LC               -0.93          1.00               -0.52               
0.52                -0.47                     0.48
xmid.(Intercept)            0.55         -0.52                1.00              
-0.93                 0.54                    -0.49
xmid.Est.LC                  -0.51          0.52               -0.93            
1.00                -0.50                     0.52
scal.(Intercept)             0.50         -0.47                0.54             
-0.50                 1.00                    -0.91
scal.Est.LC                   -0.46          0.48               -0.49           
0.52                -0.91                    1.00

The 6 by 6 correlation matrix above has an element corresponding to each of the
fixed effects (3) for each group (2) (3*2=6 total). However, I want to hold the
correlation structure for the fixed effects constant across both groups.  I am
thus interested in obtaining  a three by three correlation matrix that
corresponds to the three parameters defining each groups fixed effects.

The nlme help file discusses various correlational structures for the within
component, but I couldn't find anything on holding correlations/covariances
constant across the grouping structure for the fixed effects.

Is there anything wrong with such a correlation structure for the fixed effects?
Can it be done using the nlme package?

Any thoughts would be greatly appreciated,
Ken 

		
---------------------------------

	[[alternative HTML version deleted]]