Laurent Fanchon
2005-Apr-20 15:20 UTC
[R] Habituation model : several sequences in several sessions, should I use getGroups?
Dear all, I am looking at habituation of dogs trotting on a treadmill. Each record is made of 40 to 50 data for the same variable (for example Peak). I get each record at several minutes (1, 2 and 4) for each session. And I have 4 sessions of training (one session a week). The aim is to study the effect of the factor (Minute) on the Peak variable, and to study the changes of this effect regarding to the session. I am really new to R and stats but I am learning... I don't have any R-guru around me so maybe I am really doing it the wrong way... As I measure the same dog several times, I am doing a repeated measures so I use the nlme package. The minute factor is nested in the session factor so I use the function getGroups. Here is my script : treadmill$Time=getGroups (treadmill,form=~1|Session/Minute, level=2) Habituation <- lme (fixed=Peak~Time,data=treadmill,random=~1|Dog/Session/Minute) This seems to work as I get the random effects of Dog, Session in Dog, Minute in Session in Dog and the fixed effects of Peak~Time. The problem is in the fixed effects section, instead of listing the different times, it writes Intercept (Normal) but then Time.L, Time.Q, Time.C, Time^4, Time^5, Time^6, ..., Time^11. My treadmill$Time list contains the following : 1/1, 1/2, 1/4, 2/1, 2/2, 2/4,... corresponding to Session/Minute. I expected it names the fixed effect according to this list. So I wonder what this lines means? Is it corresponding to the data of the treadmill$Time list? Thank you very much for your help Laurent