Skip to content

lme: Specifying a formula

2 messages · Christian Hennig, Douglas Bates

#
Dear list,

I am faced with the following model:

y=E+P+M+H+PxE+Error

y is a response, E and P are factors with fixed effects.
M is a random effect nested in P and H is a random effect nested in M.
PxE is interaction of P and E.

It seems that I should fit such a model with the lme function of library
nlme, but I was not able to figure out from the help page how to specify the
formula. In the nlme documentation, the term "nested" is always associated
with a "grouping", and I do not know what my grouping is here.

Can anybody tell me how this must be specified in a call of lme?

Thanks,
Christian


***********************************************************************
Christian Hennig
University of Hamburg, Faculty of Mathematics - SPST/ZMS
 (Schwerpunkt Mathematische Statistik und Stochastische Prozesse,
 Zentrum fuer Modellierung und Simulation)
Bundesstrasse 55, D-20146 Hamburg, Germany
Tel: x40/42838 4907, privat x40/631 62 79
hennig at math.uni-hamburg.de, http://www.math.uni-hamburg.de/home/hennig/
#######################################################################
ich empfehle www.boag.de


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
#
Christian Hennig <fm3a004 at math.uni-hamburg.de> writes:
For a complicated model like this you may find it worthwhile looking
at the examples in
@Book{pinh:bate:2000,
  author =	 {Jos\'{e} C. Pinheiro and Douglas M. Bates},
  title = 	 {Mixed-Effects Models in \textsf{S} and \textsf{S-PLUS}},
  publisher = 	 {Springer},
  year = 	 2000,
  series =	 {Statistics and Computing}
}

The grouping refers to the groups in the data with which random
effects are associated.  If the levels of the M factor are distinct
for different levels of P then you can fit your model as

 lme(y ~ E * P, data = mydata, random = ~ 1 | M/H)

The last argument indicates that there will be an additive scalar
random effect for M and for H within M.

If you do not have distinct levels for M within P you can create a new
factor with 

 getGroups(~ 1 | P/M, data = mydata, level = 2)

and assign it as the grouping factor.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._