R help - Mar 2005 - Habituation model

Dear all, 

I am looking at habituation of dogs trotting on a treadmill. 
I record the ground reaction force and I analyze it with several discrete
variables (maximum, minimum,...)
For each variable, I get between 40 and 50 data per sample. 

I record data at time 1 min, 2 min, and 4 min a day, and I have 4 days of
measurement (one day a week).
That means I have 12 samples : Day1_Min1, Day1_Min2, Day1_Min4, Day2_Min1,...

Furthermore, I have 28 dogs to analyze. 

My questions is : when can I consider the data stabilized?
The problem is I am studying habituation with several sessions. 
It seems logical that values from each first minute could be very dissimilar to
others.
It is not a fully linear training system. 
I have already done some ANOVAs with "Day" and "Min"
factors. I found a significant effect.
It is quite logical because the dog is learning. The question is then when does
it stop learning? or more precisely when is it trained enough to be analyzed?

I could do comparisons among all samples with Student test, but it is surely a
simple approach.

I can presuppose the maximal allowed variability for each variable : in the
region of 5%.

I am really new to both R and stats so if these questions are very simple and I
am just missing something, suggestions about good texts or examples on R would
be great.
I am generating data with Scilab, I have a single matrix corresponding to each
dog. But I can change it if needed.

Any help would be greatly appreciated

Thanks

Laurent Fanchon
DVM, MS
Ecole Nationale Veterinaire d'Alfort
France

since you measure the same dog several times its measurements are 
correlated and you should take this into account in your analysis 
(i.e., "aov()" is not appropriate in this case). Probably you could 
find functions "lme()" and "gls()" in the "nlme"
package very useful
for your problem and for which a very good reference is:

@Book{pinheiro.bates:00,
  author    = {J. Pinheiro and D. Bates},
  title     = {Mixed-Effects Models in S and S-PLUS},
  year      = {2000},
  address   = {New York},
  publisher = {Springer-Verlag}
}


I hope it helps.

Best,
Dimitris

----
Dimitris Rizopoulos
Ph.D. Student
Biostatistical Centre
School of Public Health
Catholic University of Leuven

Address: Kapucijnenvoer 35, Leuven, Belgium
Tel: +32/16/336899
Fax: +32/16/337015
Web: http://www.med.kuleuven.ac.be/biostat/
     http://www.student.kuleuven.ac.be/~m0390867/dimitris.htm




----- Original Message ----- 
From: "Laurent Fanchon" <lfanchon at vet-alfort.fr>
To: <r-help at stat.math.ethz.ch>
Sent: Wednesday, March 30, 2005 2:06 PM
Subject: [R] Habituation model

> Dear all,
> I am looking at habituation of dogs trotting on a treadmill. I 
> record the ground reaction force and I analyze it with several 
> discrete variables (maximum, minimum,...)
> For each variable, I get between 40 and 50 data per sample.
> I record data at time 1 min, 2 min, and 4 min a day, and I have 4 
> days of measurement (one day a week). That means I have 12 samples : 
> Day1_Min1, Day1_Min2, Day1_Min4, Day2_Min1,...
>
> Furthermore, I have 28 dogs to analyze.
> My questions is : when can I consider the data stabilized?
> The problem is I am studying habituation with several sessions. It 
> seems logical that values from each first minute could be very 
> dissimilar to others. It is not a fully linear training system. I 
> have already done some ANOVAs with "Day" and "Min"
factors. I found
> a significant effect.
> It is quite logical because the dog is learning. The question is 
> then when does it stop learning? or more precisely when is it 
> trained enough to be analyzed?
>
> I could do comparisons among all samples with Student test, but it 
> is surely a simple approach.
> I can presuppose the maximal allowed variability for each variable : 
> in the region of 5%.
> I am really new to both R and stats so if these questions are very 
> simple and I am just missing something, suggestions about good texts 
> or examples on R would be great. I am generating data with Scilab, I 
> have a single matrix corresponding to each dog. But I can change it 
> if needed.
> Any help would be greatly appreciated
>
> Thanks
>
> Laurent Fanchon
> DVM, MS
> Ecole Nationale Veterinaire d'Alfort
> France
>
> ______________________________________________
> 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
>