Re: Statistics

From: Mark A. KuKuruga (kukuru@umich.edu)
Date: Fri Oct 08 1999 - 16:47:10 EST


Janko,
we're talking about using K-S test for comparing histograms . . .
I've tried to use the BD version of this test, and I agree . . . it's
not at all intuitive, and BD's no help.  So,  looking to the literature
. . .
Of course, Howard briefly discusses the virtues of K-S comparisons in
his Practical Flow book (3rd Edition, page 197).  He agrees that small
histogram differences (i.e., non-significant) will actually result in
K-S results indicating significance.  I could discuss this further, but
I suggest you go to Howard's for some insights and references.
MAK.

Nikolic-Zugic_Janko/mskcc_SU@mskmail.mskcc.org wrote:

>           Name: Re:
>    Re:    Type: Plain Text (text/plain)
>       Encoding: 7bit

--
Mark A. KuKuruga, Managing Director
University of Michigan Core Flow Cytometry
<http://www.cancer.med.umich.edu/flow_cytometry>
phone: 734-647-3216  fax: 734-936-7376
kukuru@umich.edu



This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:54:04 EST