Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: Nonparametric two way ANOVA


From   Roger Newson <roger.newson@kcl.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   st: Re: Nonparametric two way ANOVA
Date   Thu, 05 Feb 2004 20:57:46 +0000

At 14:08 05/02/04 -0500, Joseph Wagner wrote (in response to my suggestion re geometric means):

>In my experience, viral loads are lognormally distributed. I would
>therefore log-transform the viral loads and use linear regression methods
>on the logs to estimate geometric means and their ratios, using the
>-eform()- option of -regress-.


Yes, this is generally true but unfortunately in my case, log-transformation
does not help.  Is there a nonparametric method I can use?
One possibility might be the -somersd- package, downloadable from SSC, together with -parmby-, part of the -parmest- package, also downloadable from SSC. The -somersd- package calculates confidence intervals for Somers' D and Kendall's tau-a, saving the results as estimation results. The -parmby- command calls an estimation command (like -somersd-) and creates an output data set with 1 observation per parameter per by-group and data on estimates, standard errors, confidence intervals, P-values and other parameter attributes.

Joseph doesn't specify whether the really interesting predictor is race or age group, or whether race is in any way ordinal (including binary eg 1 for "Black" and 2 for "White"). However, assuming that race is a boring categorical variable and age group is an interesting ordinal variable, then Joseph might calculate a Somers' D for the association of viral load with age for each race, and then meta-analyse these Somers' D values for each race to create a grand mean Somers' D measuring the ability of higher age to predict higher viral load within races.

More information on the -somersd- and -parmest- packages is available on my website (see my signature). Somers' D and Kendall's tau-a are detailed in my Stata Journal paper (Newson, 2002), which is available in pre-publication draft form on my website. (As is my Stata tip about calculating geometric means using the -eform()- option of -regress-.)

I hope this helps. I would be able to make more specific suggestions if I knew more about precisely what association Joseph is trying to measure.

References

Newson R. Parameters behind "nonparametric" statistics: Kendall's tau, Somers' D and median differences. The Stata Journal 2002; 2(1): 45-64.


--
Roger Newson
Lecturer in Medical Statistics
Department of Public Health Sciences
King's College London
5th Floor, Capital House
42 Weston Street
London SE1 3QD
United Kingdom

Tel: 020 7848 6648 International +44 20 7848 6648
Fax: 020 7848 6620 International +44 20 7848 6620
or 020 7848 6605 International +44 20 7848 6605
Email: roger.newson@kcl.ac.uk
Website: http://www.kcl-phs.org.uk/rogernewson

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

*
* 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/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index