[R] two sided wilcox test

[Peter Ehlers]

On 2012-05-23 10:19, csrabak wrote:
> Em 23/5/2012 11:59, evaD escreveu:
>> Hey, could I just ask you a quick question?
>>
>>
>> if i have two sets of numbers, eg:
>>
>> list1 = 1,3,4,4,4 list2 = 1,2,4,5,6
>>
>>
>>
>> and in R, i run the command:
>>
>> wilcox.test(list1, list2, paired=TRUE, alternative="greater")
>>
>> and say i get a p-value of 0.0001
>>
>> Does this tell me that my values in list 1 tend to be larger than my
>> values in list 2, and that the difference in the distributions is
>> significant?
>>
> Read the third paragraph of the "Details" section of help page for this
> function, which you can reach typing:
>
>   >  ?wilcox.test

Or, if that's not clear enough, why not just try it
on a case with known result:

   y <- 1:10
   x <- y + runif(10)
   wilcox.test(x, y, ....)

Peter Ehlers