[SciPy-user] help with scipy.stats.mannwhitneyu

Bruce Southey bsouthey@gmail....
Thu Feb 5 08:56:08 CST 2009


Wavy Davy wrote:
> Hi all
>
> I am using the mannwhitneyu in the stats module, and I was looking the
> code and I see this notice in the docstring.
>
> "Use only when the n in each condition is < 20 and you have 2
> independent samples of ranks. "
>
> Am I reading it correctly that this test should only be used with
> sample sizes less than 20?
>
> I am not a statistican, more a python coder. I have been pointed and
> this test as a more robust version of the t-test, so forgive my
> ignorance.
>
> Any help would be much appreciated.
>
>   
I think the docstring is referring to the distribution of the actual 
U-test. So for small samples typically the pvalue is directly computed 
from the sampling distribution. However, Scipy is using the normal 
approximation is which is not meant to be that great.

http://faculty.vassar.edu/lowry/utest.html
http://faculty.vassar.edu/lowry/ch11a.html
http://www.alglib.net/statistics/hypothesistesting/mannwhitneyu.php

Bruce


More information about the SciPy-user mailing list