Wendell Joice wrote:
> I am contemplating running some 3-way chi-squares
> (chi-square tests on three-way tables).
One aspect of this is that often people forget that in some tables there is
a certain logic to the categories, e.g. small, larger, very large. In this
situation you have an ordinal structure for which it is simplifying too
much to use Chi2 square tests which will not reflect the ordering. More
attention should then be given to measures such as Goodman & Kruskal's
gamma coefficient
For a single table you get Gamma in Stata by:
tab v1 v2, gam
This can also be worked with like Mantel Haenzel summary of 2x2x2 tables in
tables of any dimension with the socalled partial gamma or weighted gamma
coefficient.
In Stata the procedure partgamm will do this, including giving confidence
intervals for Gamma. The normal Partgam was made by Svend Kreiner and I.
You get partgam by: findit partgam
or send me a mail and I will send it to you.
Jens M. Lauritsen, Consultant,MD. PhD.
Accident Analysis Group, Odense University Hospital,
Odense, Denmark e-mail: [email protected]
Initiator of EpiData, see www.epidata.dk for DataEntry in windows (EpiInfo
V6 principle)
Phone +45 6541 2293 Mobile +45 2173 2293 fax: +45 6613 6581
*
* 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/
