Re: st: Equivalent to kruskal-wallis in clustered data

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

Yes, there is a clustered version of the Kruskal-Wallis test. It can be 
done using the -somersd- package (downloadable from SSC) with -xi:- and 
-testparm-.

For instance, in the -auto- data, we might test independence of price 
and repair record, assuming that we are sampling car firms from a 
population of car firms, instead of sampling car models from a 
population of car models. We set up the data by typing:

sysuse auto, clear
gene firm=word(make,1)
tab firm, m

This creates and tabulates  a new variable -firm-, indicating the firm 
that makes each car model. We then do the analysis by typing:

xi, noomit: somersd price i.rep78, transf(z) tdist

which creates variables _Irep78_1 to _Irep78_5, indicating membership of 
each of the 5 repair record groups, and calculates a Somers' D of each 
of these indicators with respect to -price-, with confidence limits and 
a P-value. These Somers' D parameters measure the association of each 
repair record group (compared to all other repair record groups) with 
the car's price in dollars.

To do the test, we then type:

testparm _I*

which tests the hypothesis that all 5 of these Somers' D parameters are 
zero, which implies that no repair group tends to be more or less 
expensive than the rest (the hypothesis usually tested using a 
Kruskall-Wallis test). We see that the P-value is 0.5618, so the null 
hypothesis has not been decisively refuted.

I hope this helps. Let me know if you have any further queries.

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.

On 08/11/2012 10:11, alfonsa leiva wrote:
> Dear fellows
>
>
> Basically, GPs were randomized to 3 groups, in the bivariate analysis of effectiveness dependent variable are continuos and independet variable are groups 1,2 or 3 . There is any test equivalent to kruskall-wallis implemented in stata to test the study hypothesis adjusted for the lack of independency of the patients(clustered data by GPs)?
>
> Thanks in advance
>
>
>
> Alfonso Leiva
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/