[R-sig-eco] Question about npmc package

[David Hewitt]

> I am trying to use the npmc (non-parametric multiple comparisons) 
> package on my data set. Because there are 12 treatments, there are 66 
> post hoc comparisons.
> 

That's an awful lot of comparisons. A lot of statisticians would cringe at
what you're doing. For a comical introduction, see

http://www.blackwell-synergy.com/doi/abs/10.1111/j.1740-9713.2007.00216.x

You're virtually guaranteed a spurious result in there somewhere if you rely
on p-values for that many post-hoc tests.



> The package returns a very nice table, but I am 
> having trouble interpreting it, because the comparisons are relabeled 
> with numbers (1-2, 1-3, 1-4, etc...) instead of the labels that I've 
> given each treatment in the vector "class"
> 
> It seems as though setting the argument "info" should be able to return 
> information like the names of the compared groups, but when I try to 
> use "group.index" or when I try to set info=TRUE, I get these errors,
> 
>> npmc(data,info=group.index)
> Error in npmc(data, info = group.index) : object "group.index" not found
> 

'info' is not an argument to the npmc function, but rather an argument
passed to 'summary' when it is used on the object returned by npmc. type
?npmc at the prompt

The example in the help file indicates that this might get you what you
want:
> summary(npmc(yourdata), info=TRUE)




-----
David Hewitt
Research Fishery Biologist
USGS Klamath Falls Field Station (USA)
-- 
View this message in context: http://www.nabble.com/Question-about-npmc-package-tp18021062p18033019.html
Sent from the r-sig-ecology mailing list archive at Nabble.com.