NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Rejecting outliers: was: Kurtosis.
From: George Huxtable
Date: 2011 Jan 1, 22:51 -0000
From: George Huxtable
Date: 2011 Jan 1, 22:51 -0000
Marcel Tschudin wrote, about using the median instead of the mean-
"The median is a more "robust" estimate when only a few data are available.
In order that the median does not coincide with one of the values within
the dataset I take even numbers of observations whenever possible."
What does robust mean in this context? Taking the median is an extreme case
of rejecting outliers. It involves rejecting all the outer members of a
distribution, in pairs of highest and lowest values at a time, until only
one or two are left, depending on whether ann odd or even number of values
were taken. In Marcel's case, two, it seems.
But starting with a normal distribution, the resulting scatter in the
median is significantly greater than the scatter in the mean, simply
because so much of the numerical information has been discarded. So Marcel
is paying a heavy price. I wonder if he has made an analysis of any
benefits of this procedure.
George
contact George Huxtable, at george{at}hux.me.uk
or at +44 1865 820222 (from UK, 01865 820222)
or at 1 Sandy Lane, Southmoor, Abingdon, Oxon OX13 5HX, UK.
