RE: st: Wilcoxon Matched-Pairs Signed-ranks Test also for groups?

["Newson, Roger B" <r.newson@imperial.ac.uk>]

If you wanted a confidence interval for the median difference between the 2 groups, then you can use the -cendif- module of the -somersd- package (which you can download from SSC(, with a cluster() option.

The -somersd- package comes with 3 .pdf manuals, which you can copy to your system during installation. Example 3.2 in cendif.pdf is similar to your example. There are 2 median differences to choose from, namely the pairwise median difference (the median pairwise difference between the 2 matched member of a pair) and the group median difference (ie the median difference between the Group A member of one randomly-chosen pair and the Group B member of another randomy-chosen pair). The first of these is estimated using the option -funtype(wclust)-, and usually has a wider confidence interval. The second of these is estimated using the option -funtype(vonmises)-, and usually has a narrower confidence interval.

I hope this helps.

Best wishes

Roger


Roger B Newson BSc MSc DPhil
Lecturer in Medical Statistics
Respiratory Epidemiology and Public Health Group
National Heart and Lung Institute
Imperial College London
Royal Brompton Campus
Room 33, Emmanuel Kaye Building
1B Manresa Road
London SW3 6LR
UNITED KINGDOM
Tel: +44 (0)20 7352 8121 ext 3381
Fax: +44 (0)20 7351 8322
Email: r.newson@imperial.ac.uk 
Web page: http://www.imperial.ac.uk/nhli/r.newson/
Departmental Web page:
http://www1.imperial.ac.uk/medicine/about/divisions/nhli/respiration/popgenetics/reph/

Opinions expressed are those of the author, not of the institution.

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of jverkuilen
Sent: 30 January 2009 06:57
To: statalist@hsphsun2.harvard.edu
Subject: RE: st: Wilcoxon Matched-Pairs Signed-ranks Test also for groups?

See -reshape-.  

However, I am not entirely convinced you want to run a non-par test. What is the nature of the non-normality? Most non-par tests lose the original scale of the variable, which isn't nice. Also they are badly affected by ties. The paired t test might still be good, so long as you don't have marked skew. 

JV

-----Original Message-----
From: "Rico Baumann" <Rico.Baumann@gmx.de>
To: statalist@hsphsun2.harvard.edu
Sent: 1/29/2009 8:41 PM
Subject: st: Wilcoxon Matched-Pairs Signed-ranks Test also for groups?

Hi,

I have data with a binary variable (0 and 1) that splits my data set into two matched groups (one group of original observations and one control group), therefore the samples are not independent. 

group1  0  varname1 ...
group2  1  varname1 ...

Since some of the contained variables are not normally distributed I´d like to use a non-parametric test in order to test for the median difference of some variables between the two dependent groups.

The above mentioned signrank -- Equality tests on matched data seems to be the correct test, but unfortunately the syntax doesn´t allow the option by(grouppvar), like e.g. ttest or ranksum do.
  signrank varname = exp [if] [in]

Is there any (advanced) solution for that, or do I have to rearrange the data in a spreadsheet outside Stata?

Thanks for help and best, 
Rico
-- 
Beste Grüße
Rico Baumann

email:   rico.baumann@gmx.de
cell:      +49 (0)172 532 25 58
phone:  +49 (0)211 239 40 500  
efax:     +49 (0)3212 - 1077734
address: Vinckestraße 5, 40470 Düsseldorf, 
             Deutschland / Germany

  
 
 
 







*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/