Understanding the Wilcoxon-signed Rank test

  • Thread starter Thread starter chwala
  • Start date Start date
  • Tags Tags
    rank Test
chwala
Gold Member
Messages
2,825
Reaction score
413
Homework Statement
I would like insight on the highlighted part in red.
Relevant Equations
stats
Reference; https://en.wikipedia.org/wiki/Wilcoxon_signed-rank_test#:~:text=The Wilcoxon signed-rank test,-sample Student's t-test.

I have managed to go through the literature (it is pretty straight forward). In general Wilcoxon-rank method applies to data with unequal variances otherwise student t- test would be sufficient.

Now looking at the signed ranks given (check attachment) we have;

##[1,2,3,4,5,6,7,8,9,10,11,-12]##

In my understanding, we shall have ##W^{+} =66, W^{-}=12## and we also know that,

##\dfrac {n(n+1)}{2}=\dfrac {12(13)}{2}=78=[W^{+}+W^{-}]##

therefore it follows that, test statistic ##W=12##.

Now to my question, how did they arrive at p-value ##\left[\dfrac{55}{2^{12}}\right]##?

Secondly, how did they arrive at the given signed-ranks without the 'ordered absolute value of differences'?

thanks...
 

Attachments

Last edited:
Physics news on Phys.org
I think i got it; The p-value is given by;

##P(W=w)=f(w)= \dfrac{c(w)}{2^n}## where ##c(w)## is the number of possible ways to assign a+ or a- to the first n integers. For our case we have;

##c(w)=\dfrac{n(n+1)}{2}=\dfrac {10×11}{2}=55##.

Therefore,

##P(W=w)=f(w)= \dfrac{55}{2^{12}}##
 
Note that the t-test can be used (with some modification) for unequal variances. The point about the Wilcoxon test is that it is applicable to data that are not normally distributed, where the t-test assumes normality.

The problem with the uneqal variance test was neatly summed up in (I think) Cochran's book on sample design: a hundred men try a hair restorer product. Fifty go bald and fifty turn into werewolves. Blindly applying the inequal variance test to compare them to a control is just asking "but, were they hairier on average?"
 
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...
Back
Top