Often it is easier to understand
precisely what is going on by
looking at the code. In your case
. findfile ranksum.ado
. doedit `r(fn)'
shows that Prob > |Z| is,
in Stata language and in this case,
2 * normprob(-0.354)
or
2 * (1 - normprob(0.354))
In other words, it is a P-value
calculated on the assumption that
under the null hypothesis
the sampling distribution of z is Gaussian
with mean 0 and variance 1. Note that
sign is disregarded.
With -ranksum- I always like to see the
results of the -porder- option, which has
a substantive interpretation.
Nick
n.j.cox@durham.ac.uk
Andreas Aschbacher
> I made two sample Wilcoxon-ranksum-test
> / to show that two independent samples are from same
> distribution = H0/
> with following results: m ... men W ... woman
> -
> gender | obs rank sum expected
> -------------+---------------------------------
> m | 3.7e+05 1.397e+11 1.398e+11
> W | 3.7e+05 1.382e+11 1.382e+11
> -------------+---------------------------------
> combined | 7.5e+05 2.780e+11 2.780e+11
>
> unadjusted variance 8.636e+15
> adjustment for ties -1.316e+08
> ----------
> adjusted variance 8.636e+15
>
> Ho: dosis(gender==m) = dosis(gender==W)
> z = -0.354
> Prob > |z| = 0.7236
>
>
> -
>
> because of theorie in my books and referencebook in Stata is not
> "exactly the same" I want to assure about denotation of results ?
> z= -0.354 is check-value - o.k.
> but I don't understand : Prob > |z| = 0.7236
> Confidence-intervall that H0 is right - is ~72% ??
> but what does mean : Prob > |z| ?
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/