R help - Mar 2006 - nlme for groupedData with inner and outer factors

Hello,

I am having trouble specifying a suitable nlme model.
My data structure is described by

gd <- groupedData(ppath ~ lcut | exp, outer = ~ bait, inner = ~ weight, data 
= d)

i.e. the response (ppath) of several subjects (sub) was measured at levels 
of a continuous variable (lcut). Subjects were given either of one level of 
a factor (bait), and all subjects were measured at two levels of another 
factor (weight). Therefore bait varies among subjects and weight varies 
within subjects.

The relationship ppath ~ cut for each subject and weight appear to follow a 
logistic curve, with xmid and scal affected by bait and weight. There is 
also a random effect of subject on xmid and scal.

Any help with formulating the correct model would be greatly appreciated.

Many thanks,
Dan Bebber

Department of Plant Sciences
University of Oxford

p.s. Part of my data are shown below:

     sublcut       ppath bait weight
1   pv1_  0.0 1.000000000    1      0
2   pv1_  0.1 0.827773821    1      0
3   pv1_  0.2 0.380102502    1      0
4   pv1_  0.3 0.209151878    1      0
5   pv1_  0.4 0.076929304    1      0
6   pv1_  0.5 0.065681564    1      0
7   pv1_  0.6 0.020670108    1      0
8   pv1_  0.7 0.012817021    1      0
9   pv1_  0.8 0.008661514    1      0
10  pv1_  0.9 0.011568323    1      0
11  pv19  0.0 1.000000000    1      0
12  pv19  0.1 0.668390291    1      0
13  pv19  0.2 0.343318462    1      0

1.  Have you read Pinheiro and Bates (2000) Mixed-Effects Models in S 
and S-Plus (Springer)?  If no, I believe your study of that book will be 
well rewarded;  mine has.

	  2.  If you've looked at Pinheiro and Bates and still have questions 
about this, PLEASE do read the posting guide! 
"www.R-project.org/posting-guide.html", especially the bit about 
developing a toy example that is as simple as you can make it that still 
illustrates your question.  I've solved many of my own problems doing 
this, and I've answered many questions for people on this list dealing 
with functions I've not previously used.  You could help people like me 
by providing a few lines of R code that we could copy from your email, 
paste into R and replicate what you see.

	  hope this helps,
	  spencer graves

Dan Bebber wrote:
> Hello,
> 
> I am having trouble specifying a suitable nlme model.
> My data structure is described by
> 
> gd <- groupedData(ppath ~ lcut | exp, outer = ~ bait, inner = ~ weight,
data
> = d)
> 
> i.e. the response (ppath) of several subjects (sub) was measured at levels 
> of a continuous variable (lcut). Subjects were given either of one level of
> a factor (bait), and all subjects were measured at two levels of another 
> factor (weight). Therefore bait varies among subjects and weight varies 
> within subjects.
> 
> The relationship ppath ~ cut for each subject and weight appear to follow a
> logistic curve, with xmid and scal affected by bait and weight. There is 
> also a random effect of subject on xmid and scal.
> 
> Any help with formulating the correct model would be greatly appreciated.
> 
> Many thanks,
> Dan Bebber
> 
> Department of Plant Sciences
> University of Oxford
> 
> p.s. Part of my data are shown below:
> 
>      sublcut       ppath bait weight
> 1   pv1_  0.0 1.000000000    1      0
> 2   pv1_  0.1 0.827773821    1      0
> 3   pv1_  0.2 0.380102502    1      0
> 4   pv1_  0.3 0.209151878    1      0
> 5   pv1_  0.4 0.076929304    1      0
> 6   pv1_  0.5 0.065681564    1      0
> 7   pv1_  0.6 0.020670108    1      0
> 8   pv1_  0.7 0.012817021    1      0
> 9   pv1_  0.8 0.008661514    1      0
> 10  pv1_  0.9 0.011568323    1      0
> 11  pv19  0.0 1.000000000    1      0
> 12  pv19  0.1 0.668390291    1      0
> 13  pv19  0.2 0.343318462    1      0
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html