Re: [R] value of W seems to be suspicious in the mann-whitney wilcox related test. what could be the problem

[taby gathoni]

Dear Abhijit,

Maybe am using the wrong test. But all i need is calculate non-parametrically 
the accuracy ratios of my samples.
Ideally for that kind of an output then the answer would be between 0 and 1.
notice also one of my variable is a non-numeric.

Thank
Taby









An idea not coupled with action will never get any bigger than the brain cell 
it occupied.
Arnold Glasgow
......
"Attempt something large enough that failure is guaranteed…unless God steps 
in!"


--- On Mon, 3/14/11, Abhijit Dasgupta <aikidasgu...@gmail.com> wrote:

From: Abhijit Dasgupta <aikidasgu...@gmail.com>
Subject: Re: [R] value of W seems to be suspicious in the mann-whitney wilcox 
related test. what could be the problem
To: 
Cc: tab...@yahoo.com, "andrija djurovic" <djandr...@gmail.com>, "R help" 
<r-help@r-project.org>
Date: Monday, March 14, 2011, 5:16 PM



  

    
  You need to read up on the Wilcoxon signed-rank test and the output
    from wilcox.test!!!

    

    The confidence interval is of the difference of medians, which can
    certainly be negative. In fact, your estimate is -33, and the
    confidence interval is (-68, 0) which is reasonable.

    

    The value of W is a positive number, in general, and isn't
    restricted to [0,1]

    

    Abhijit

     

      On 3/14/2011 11:07 AM, taby gathoni wrote:
      
        my output is as follows:
 wilcox.test(main_samp$SCORE~main_samp$GENDER, conf.int = TRUE)

        Wilcoxon rank sum test with continuity correction

data:  main_samp$SCORE by main_samp$GENDER
W = 2780.5, p-value = 0.04829
alternative hypothesis: true location shift is not equal to 0 
95 percent confidence interval:
 -6.800005e+01 -2.056837e-05 
sample estimates:
difference in location 
             -33.00003 
 result of W seems suspicious since i expect the result to be between 0 and 1. 

and the confidence intervals are also -ves  what could be the challenge?
Thanks Taby







      
        [[alternative HTML version deleted]]


        
______________________________________________
R-help@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.

      
    
  



      
        [[alternative HTML version deleted]]


______________________________________________
R-help@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.