Ricardo Ovaldia wrote (excerpted):
My doubt arose because I get a highly statistical difference where I was not
expecting one. The mean from the two culture mediums are similar (9.2 vs. 10.2)
and the medians and rage are exactly the same:
N Mean (SD) Median Min - Max
-----------------------------------------------------
Percent_0 7010 9.2 (8.2) 10.0 0.0 - 50.0
Percent_1 7390 10.2 (9.2) 10.0 0.0 - 50.0
-----------------------------------------------------
However, -vanelteren- yields p=0.0054 (asymptotic) p=0.002 (permutation based)
and that seems strange:
. vanelteren percent, by(mm) strata(dam)
Generalized Wilcoxon-Mann-Whitney Ranksum Test (van Elteren's Test)
Variance of
Weighted Expected Weighted
Stratum | n | Ranksum | Ranksum | Ranksum
-------------+------+----------+---------+------------
1 | 6 | 1.00 | 1.0 | 0.068
2 | 2 | 0.50 | 0.5 | 0.000
3 | 4 | 1.00 | 1.0 | 0.000
4 | 3 | 1.00 | 1.0 | 0.042
.
.
.
234 | 6 | 1.57 | 1.0 | 0.087
235 | 9 | 1.40 | 2.0 | 0.113
236 | 6 | 0.86 | 1.0 | 0.084
-------------+------+----------+---------+------------
Sums |1,485 | 375.02 | 363.0 | 18.646
--------------------------------------------------------------------------------
That the medians of the pooled data are identical wouldn't bother me so much as
the difference between the asymptotic and permutation p-values with 256 dams.
Take a look at what -xtreg percent mm, i(dam) fe- gives you (and also take a
look at, say, -pnorm- on the residuals, for starters). I'm guessing that
-xtreg, fe- (which would have been my first choice with this design) gives you
the same take-home message as what -vaneltern- does.
Joseph Coveney
*
* 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/
