[R] How to do multiple comparisons with multiple error strata

[Torsten Hothorn]

> Hello,
>   the title says it all. I have an ANOVA model with
> "Error" term and the functions TukeyHSD or those
> from the "multcomp" package seem not to work with
> such a model. Assuming that the object returned from
> aov() function with multiple strata is a list of aov objects,
> I tried something like TukeyHSD(aov.1$Within), but this
> does not work either.
>  Does somebody have a solution? I think I can calculate
> confidence interval for individual coefficients, extending
> interval ranges based on Bonferroni correction, but this
> seems to be too crude solution...

as far as I understand this, an aov model is fitted for each error
stratum. If you are able to compute the correlation between the
(least square) estimators you compute the contrasts from, 
you can use `csim{test,int}' by directly
specifying a) estimators b) correlation matrix and c) contrast matrix
(which you can compute with `contrMat' for Tukey contrasts). 

Torsten

> 
> Thank you,        Petr Smilauer
>  
> 
> 
> -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> 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
> _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> 

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._