[R] Wilcoxon signed-ranks test using package coin ?

Peter Ehlers ehlers at ucalgary.ca
Tue Feb 9 03:49:44 CET 2010


Dale Steele wrote:
> Given the following data, and hypothesized median M.0 I've found a
> method to implement the Wilcoxon signed-rank test.
> 
> Data: (with one zero difference and tied ranks)
> 
> x <- c(136, 103, 91, 122, 96, 145, 140, 138, 126, 120, 99, 125,
> 91,142, 119, 137)
> M.0 <- 119
> 
>> library(exactRankTests)
>   Package ‘exactRankTests’ is no longer under development.
>   Please consider using package ‘coin’ instead.
> 
>> wilcox.exact(x, mu=M.0)
> 
> 	Exact Wilcoxon signed rank test
> 
> data:  x
> V = 65.5, p-value = 0.771
> alternative hypothesis: true mu is not equal to 119
> 
> 
> I've been unable to implement this test using library(coin) - is this possible?
> 
> Thanks.  --Dale
> 
Try this:

  wilcoxsign_test(x ~ rep(M.0, length(x)), dist = 'exact')

-- 
Peter Ehlers
University of Calgary



More information about the R-help mailing list