Re: [S] distribution-free tests for dispersion

Anthony Rossini (rossini@math.sc.edu)
26 Mar 1998 10:03:57 -0500


>>>>> S Raessler <Susanne.Raessler@wiso.uni-erlangen.de> writes:

>does anyone have some S-Plus-functions of test statistics which could
>be used, when the assumption of normal distribution is invalid and a
>test of identical variances shall be performed. Unfortunately our
>distributions observed are well skewed to the right. Thus we are
>searching for tests for dispersion in the two-sample case such like
>the Siegel-Tukey, Mood, Ansary-Bradley, Moses or similar ones.

They are not too hard to write up. See some of the (incomplete)
implementation somewhere under
http://franz.stat.wisc.edu/~rossini/couirses/intro-nonpar/

also, on the same machine, is the ctest package, in the R archive
CRAN, which ought to work (maybe with small modifications) under S.
Start at
http://franz.stat.wisc.edu/
and look for the CRAN packages (somewhere in CRAN).

best,
-tony

--
Anthony Rossini
Manifold Graphics / Epimetrics

----------------------------------------------------------------------- This message was distributed by s-news@wubios.wustl.edu. To unsubscribe send e-mail to s-news-request@wubios.wustl.edu with the BODY of the message: unsubscribe s-news