R help - Jan 2007 - nested hierarchical design

Dear R-Helpers,

I would like to know what syntax I need to use to do a nested anova for 
1. a continuous variable and 2. count data (x out of y)
1. The first I used to do in SPSS and I would like to be able to do it 
in R as well.
This is the hierarchical model I would like to use: a continuous 
variable explained by factor A(fixed) + factor B(random) nested in A + 
factor C (random) nested in factor B (which is nested in A).
The data is unbalanced at all levels, meaning different nrs of B per A, 
different nrs of C for B, different nrs of values per C
In SPSS this seems to be quite simple with a small writing in the syntax 
(see below). But how do I get the same results in R with lme?

SPSS syntax:
UNIANOVA
  var  BY A B C
  /RANDOM = B C
  /METHOD = SSTYPE(3)
  /INTERCEPT = INCLUDE
  /CRITERIA = ALPHA(.05)
  /DESIGN = A B(A) C(B(A))).

2. the same model but than for count data (like 15 out of 30, 23 out of 
60) instead of the continous variable(I know the basics of glm in R)

Thanks a lot for your help
Jelmer

> 1. The first I used to do in SPSS and I would like to be able 
> to do it in R as well.
> This is the hierarchical model I would like to use: a continuous 
> variable explained by factor A(fixed) + factor B(random) 
> nested in A + factor C (random) nested in factor B (which is nested
> in A).
You would need the following, I guess:

lme (var ~ A, data= NameOfDataFrame, random= ~ 1 | B/C)

This means that C is nested in B. If all C within B have the same levels of A,
the model automatically treats B as nested in A(but you can check that based on
the degrees of freedom in the output).
> 2. the same model but than for count data (like 15 out of 30, 
> 23 out of 60) instead of the continous variable(I know the basics
> of glm in R)
I guess, this can be viewed as a data set with binary outcomes and thus modeled
by a generalised mixed-effects model with a binomial distribution using glmmPQL
(in library MASS) or lmer (in library lme4). Please refer to:

@Book{Pin:00a,
  author = {Pinheiro, Jose C and Bates, Douglas M},
  title = {Mixed-Effects Models in {S} and {S}-{P}{L}{U}{S}},
  publisher = {Springer},
  year = {2000},
  address = {New York}
}

@BOOK {Ven:02,
  AUTHOR = {Venables, W N and Ripley, B D},
  TITLE = {Modern Applied Statistics with {S}},
  PUBLISHER = {Springer},
  YEAR = {2002},
  ADDRESS = {New York},
  EDITION = {fourth}
}

@Article{Rnews:Bates:2005,
  author       = {Douglas Bates},
  title	       = {Fitting Linear Mixed Models in {R}},
  journal      = {R News},
  year	       = 2005,
  volume       = 5,
  number       = 1,
  pages	       = {27--30},
  month	       = {May},
  url	       = {http://CRAN.R-project.org/doc/Rnews/},
}

and: http://wiki.r-project.org/rwiki/doku.php?id=guides:lmer-tests

Do not be surprised if the results are not identical to those in SPSS ...

Regards, Lorenz Gygax
- 
Swiss Federal Veterinary Office
Centre for proper housing of ruminants and pigs
Agroscope Reckenholz-T?nikon Research Station ART