You may want to look at the interaction function (a quick way to make the single
factor with 4 levels that you mention).
You can create your own sets of contrasts and set them using the C or contrasts
functions, then use the split argument to summary.aov to look at the individual
degrees of freedom.
You may also be interested in the multcomp package for looking at the
comparisons.
Hope this helps,
--
Gregory (Greg) L. Snow Ph.D.
Statistical Data Center
Intermountain Healthcare
greg.snow at intermountainmail.org
(801) 408-8111
> -----Original Message-----
> From: r-help-bounces at stat.math.ethz.ch
> [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of
> Ronaldo Reis Junior
> Sent: Monday, July 23, 2007 4:05 PM
> To: R-Help
> Subject: [R] Function to separate effect in AOV
>
> Hi,
>
> I have a dummy question.
>
> Suppose that I have two explanatory variable, T1 (A, B) and
> T2 (C, D) and one response variable.
>
> > attach(dados)
>
> > tapply(Y,list(T1,T2),mean)
> C D
> A 2.200000 10.20000
> B 2.223333 20.26667
>
> In this case, "A" and "B" inside "C" have no
difference, but
> have differences inside "D"
>
> I make this model:
>
> > m <- aov(Y~T1*T2)
> >
> > summary(m)
> Df Sum Sq Mean Sq F value Pr(>F)
> T1 1 76.36 76.36 5617.9 1.119e-12 ***
> T2 1 508.69 508.69 37426.7 5.704e-16 ***
> T1:T2 1 75.65 75.65 5566.0 1.161e-12 ***
> Residuals 8 0.11 0.01
> ---
> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05
'.' 0.1 ' ' 1
>
>
> This result don't show the reality of the data, because I
> cant see that "A"
> and "B" inside "C" are the same.
>
> The anova result is the same of an full different levels, like this:
>
> > attach(dados2)
> >
> > tapply(Y,list(T1,T2),mean)
> C D
> A 6.100000 10.20000
> B 2.223333 20.26667
> >
> > m <- aov(Y~T1*T2)
> >
> > summary(m)
> Df Sum Sq Mean Sq F value Pr(>F)
> T1 1 28.74 28.74 2114.3 5.529e-11 ***
> T2 1 367.75 367.75 27056.7 2.088e-15 ***
> T1:T2 1 145.81 145.81 10728.1 8.433e-14 ***
> Residuals 8 0.11 0.01
> ---
> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05
'.' 0.1 ' ' 1
>
> In this case all level are different, C to D and A to B.
>
> The question is:
>
> The only way to find this real difference is:
>
> 1) make T1 and T2 like a Treatment variable with 4 levels
> (AC,BC,AD,BD)?
>
> or
>
> 2) make 3 anova:
> a) Anova (A,B) inside C
> b) Anova (A,B) inside D
> c) Full factorial Anova (like this in the e-mail)
>
> or
>
> 3) exist any other way to make this in only one analysis, to
> find all differences e interactions? In other words, to find
> differences in "A"
> and "B" inside "C", "A" and "B"
inside "D", "C" and "D"
> inside "A" and "C"
> and "D" inside "B"
>
> Thanks
> Ronaldo
> --
> > Prof. Ronaldo Reis J?nior
> | .''`. UNIMONTES/Depto. Biologia Geral/Lab. de Ecologia
> | : :' : Campus Universit?rio Prof. Darcy Ribeiro, Vila Mauric?ia `.
> | `'` CP: 126, CEP: 39401-089, Montes Claros - MG - Brasil
> | `- Fone: (38) 3229-8187 | ronaldo.reis at unimontes.br |
> | chrysopa at gmail.com http://www.ppgcb.unimontes.br/ | ICQ#: 5692561 |
> | LinuxUser#: 205366
>
> ______________________________________________
> 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
> and provide commented, minimal, self-contained, reproducible code.
>