on 23/08/2004 15:31, Federico Soldani at [email protected] wrote:
> I�was calculating if two groups are different�for a continuous variable;
> then I realized that the distributions are quite skewed, so�I though that
> using medians and percentiles�instead of means and SDs should be more
> informative and accurate.
Try
STB-52 sg123 . . Hodges-Lehmann est. of a shift in location between two
pop.
(help npshift if installed) . . . . . . . . . . . . . . . . . D.
Wang
11/99 pp.52--53; STB Reprints Vol 9, pp.255--257
provides the Hodges-Lehmann nonparametric test (an extension
of the Wilcoxon-Mann-Whitney test) of shift between two
populations
This measure is sometimes used in clinical trials where the outcome is not
on such a scale
I also like the Mann Whitney statistic, reported by -ranksum2-
This statistic is "the probability that a randomly selected member
of one group will have a better result than a randomly selected member of
the other group"). This makes a lot of real life sense. It corresponds to
the question of 'what are the chances that someone would be better off in
one group rather than the other?"
Finally, you could try -qreg- and model the difference in the medians and
its confidence interval as a function of group membership (code the groups
as 1 and 0).
Ronan M Conroy ([email protected])
Lecturer in Biostatistics
Royal College of Surgeons
Dublin 2, Ireland
+353 1 402 2431 (fax 2764)
--------------------
Just say no to drug reps
http://www.nofreelunch.org/
*
* 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/
