[R] Mann Whitney test

[Frank E Harrell Jr]

Daniel Moreira wrote:
> Rodrigo,
> 
> Wilcoxon?Mann?Whitney allows arbitrary sample sizes. It is an extension of 
> the original Wilcoxon test which only handles equal sample sizes.
> 
> ________________________________________
> Daniel Moreira, MD

I believe you'll find that the Wilcoxon 2-sample test always allowed for 
unequal sample sizes.  To be clear there is the Wilcoxon 2-sample test 
(Wilcoxon-Mann-Whitney 2-sample rank-sum test, to be precise) and the 
Wilcoxon signed-rank one sample test, usually used for paired data (in 
which the sample sizes are equal by definition since with paired data we 
subtract a f(pre) from f(post) where f is a suitable transformation for 
the pre and post measurements from the same subject).

Frank
> 
> Research Associate
> Duke University Medical Center
> DUMC 2626, MSRB-I Room 455
> 571 Research Drive
> Durham, North Carolina 27710
> Telephone: (919) 681-7132
> Fax: (919) 668-7093
> E-mail: daniel.moreira at duke.edu
> 
> 
> 
> 
> Rodrigo Fernandes Ramalho <rofera at ig.com.br> 
> Sent by: r-help-bounces at r-project.org
> 02/14/2009 08:52 AM
> 
> To
> r-help at r-project.org
> cc
> 
> Subject
> [R] Mann Whitney test
> 
> 
> 
> 
> 
> 
> Hi,
> 
> I have a statistical question.
> 
> To apply the Mann Whitney test to compare a pair of vectors, they must to
> have the same length ?
> 
> Thanks,
> 
> Rodrigo
> 
>                  [[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org 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.
> 
-- 
Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University