R help - Sep 2009 - Non-parametric test for location with two unpaired sets of data measured on ordinal scale.

Please forgive a stats question.

I have to sets of data (unpaired) measured on an ordinal scale. I want to test
to see if the two sets are different (i.e. do they have the same location):

set1: 1,3,2,2,4,3,3,2,2 
set:   4,4,4,3,3,5,4,4

What is the most appropriate non-parametric test to test location?

Thanks,
John  

Confidentiality Statement:
This email message, including any attachments, is for th...{{dropped:6}}

What do you mean by location?  I can think of examples where 2 distributions
have the same median but different means, or the same means but different
medians.

Are you willing to assume that the distributions are exactly the same under the
null hypothesis? (not just the same 'center/location')

I would probably do a permutation test on the difference between the means or
medians (which ever you think is more meaningful), this assumes the exact same
distribution under the null.

You can also do a Mann-Whitney/Wilcoxin test (but I don't like explaining,
or sometimes even thinking about, what it is actually testing), you could do a
bootstrap confidence interval on the difference between means/medians (does not
assume distributions are the same, just have same mean/median), or you could
just replace all values by their ranks and do a t-test (essentially transforms
the data to a uniform distribution, the CLT for the uniform kicks in around n=5,
but I would simulate just to check).

This is not the nice simple answer that you were probably looking for, but
hopefully it gives you some things to think about that will help,

-- 
Gregory (Greg) L. Snow Ph.D.
Statistical Data Center
Intermountain Healthcare
greg.snow at imail.org
801.408.8111

> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
> project.org] On Behalf Of John Sorkin
> Sent: Thursday, September 24, 2009 1:08 PM
> To: r-help at r-project.org
> Subject: [R] Non-parametric test for location with two unpaired sets of
> data measured on ordinal scale.
> 
> Please forgive a stats question.
> 
> I have to sets of data (unpaired) measured on an ordinal scale. I want
> to test to see if the two sets are different (i.e. do they have the
> same location):
> 
> set1: 1,3,2,2,4,3,3,2,2
> set:   4,4,4,3,3,5,4,4
> 
> What is the most appropriate non-parametric test to test location?
> 
> Thanks,
> John
> 
> Confidentiality Statement:
> This email message, including any attachments, is for
> th...{{dropped:6}}
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-
> guide.html
> and provide commented, minimal, self-contained, reproducible code.