Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: How to analyze paired ordinal data (before and after disease) with Stata


From   Nick Cox <n.j.cox@durham.ac.uk>
To   "'statalist@hsphsun2.harvard.edu'" <statalist@hsphsun2.harvard.edu>
Subject   st: RE: How to analyze paired ordinal data (before and after disease) with Stata
Date   Fri, 27 May 2011 13:59:19 +0100

I'd check out Roger Newson's -somersd- and its relatives. -findit somersd- for locations. He will probably be posting soon about possibilities. 

But setting that aside, any absence of precedents allows you to be creative. You could score overall change in symptoms, possibly applying different weightings, and get a P-value for such change by simulation, and also apply sensitivity analysis. 

On a secondary level, -tabplot- from SSC can help with simple visualizations. See the last example in the help for a kappa-type example. 

Nick 
n.j.cox@durham.ac.uk 

Fredrik Norström

I have observed frequency of symptoms, i.e. headache, before and after diagnosis with 5 response alternatives from "never" to "always". I want to test if things have significantly improved after diagnosis. Going through the literature I notice that very few tests are suited for matched ordinal data.

I have looked at the Stata command "Symmetry" and the options for it. My judgment from the limited help file is that options within symmetry only tests symmetry and not the direction of the change. My interest is not to test if responders changed their state which is the case for e.g. Stuart-Maxwell.

Is there any function in Stata that actually tests ordinal paired data for improvement in states? If not can anyone recommend literature or test within other software so that I can test my data without using Stata? My best solution at the moment is to dichomotize my data and then it is not difficult to find a suitable test but feels like a waste to not use my data to its full extent.

Any suggestion is very much appreciated.


*
*   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/


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