s-news
[Top] [All Lists]

Extracting means from split plot models

To: s-news@lists.biostat.wustl.edu
Subject: Extracting means from split plot models
From: Mark.Hearnden@nt.gov.au
Date: Thu, 3 Apr 2008 07:26:24 +0930

Hello

I have a model which is a split plot in time, 'Sample' is nested in 'Treatment'
and repeated over 5 occasions ('Time').
It is a totally confounded model (not my idea but unavoidable) as there is only
1 replication of Treatment per year (hence the pseudoreplicated Sample level
within Treatment). Samples are slightly unbalanced within Treatment but occur
across all Times.

> splitmodel <-
aov(depvar~Treatment*Time+Error(Sample%in%Treatment+Sample%in%Treatment:Time),data=data)

My question is how do I extract the means (and se's if possible) predicted by
the model from the aov object given I have two error strata? I have tried
model.tables.aov and '.aovlist.

> model.tables.aovlist(splitmodel, se=F, type="means")

I realise I won't get se's with the unbalanced design but someone might have a
fix?

Many thanks if you can help.

(Using S+8 in WinXP)



Dr Mark Hearnden
Biometrician
Primary Industry Services
Goff Letts Building
Berrimah

Department of Primary Industry, Fisheries and Mines
Northern Territory Government of Australia
GPO Box 3000
DARWIN  NT  0801


"Everything should be made as simple as possible,
  but not simpler." Albert Einstein.


<Prev in Thread] Current Thread [Next in Thread>
  • Extracting means from split plot models, Mark . Hearnden <=