[R] post-hoc test with kruskal.test()

[David Winsemius]

There is a post hoc test along the lines of the Kruskal-Wallis test.  
It is implemented on the help page of oneway_test from package coin.  
The authors of the package, Hothorn, Hornik, van de Wiel, and  
Zeileis,  cite  Hollander and Wolfe (1999) for details and say it is  
called the NemenyiDDamico-Wolfe-Dunn test.

Or see nparcomp function in package nparcomp.

There is also a post hoc test for the situation where a Friedman test  
has been done, and that is seen on the help page for SymmetryTests in  
package coin:  the Wilcoxon-Nemenyi-McDonald-Thompson test:

http://finzi.psych.upenn.edu/R/library/coin/html/SymmetryTests.html

There is also an option of using the MTP function in the multtest  
package.

http://finzi.psych.upenn.edu/R/library/multtest/html/MTP.html

-- 
David Winsemius


On Oct 14, 2009, at 3:17 AM, Robert Kalicki wrote:

> Dear R users,
>
> I would like to know if there is a way in R to execute a post-hoc test
> (factor levels comparison, like Tukey for ANOVA) of a non-parametric
> analysis of variance with kruskal.test() function. I am comparing  
> three
> different groups. The preliminary analysis using the kruskal-wallis- 
> test
> show significance, but I still don't know the relationship and the
> significance level between each group?
>
> Do you have any suggestion?
--

David Winsemius, MD
Heritage Laboratories
West Hartford, CT