topic Re: Wilcoxon in Statistical Procedures

Wilcoxon

deleted_user — Sun, 30 May 2010 06:03:01 GMT

Hello,

I have the following data set:
Date P25 Median P75
2000 .... .... ....
2001
2002
...

I need to test the hypothesis that the median is not equal to 1.
I used the following code:

proc npar1way data = test wilcoxon;
class date;
var ... ;
run;

I receive a table looking like this:
date n Sum of Squares Expected under H0 Std under H0 Mean Score
2000 .....
2001 ....
2002 ....

At the end, it shows the Kruscal-Wallis test with one p-value.

But I need p-values for the test that the median is not equal to 1 for every year.

What do I have to change in the SAS code? Or do I have to use a different code to get the result?

Thanks,
Tina

Re: Wilcoxon

Ksharp — Sun, 30 May 2010 14:22:50 GMT

I think Wilcoxon test is a alternative way to t test.
So need class variable has two levels .
however , Your class variable has three levels .

Re: Wilcoxon

Doc_Duke — Tue, 01 Jun 2010 13:12:07 GMT

If your data are coming to you summarized as 25, 50, 75 percentiles, then you cannot do the test that you want.