MAD() on small data sets

[CdeMills]

Hello,

I'm trying to find discrepancies in small data sets, and use to this end the
mad() from package statistics. The problem is that when the number of data
is small and odd,
1) the median is one of the data
2) one of the difference with regard to the median is zero, this way the
standard deviation of the absolute deviations with respect to median and the
median of those absolute differences may be very different.

For instance, consider the sequence [496 286 292]; the median is 292, the
absolute differences are 
[ 204 6 0]; std is 116 and median is 6. What I would achieve is to spot 496
as suspicious and get as robust median of the sequence 289. An operator
where this effect is milder is 
tmp = abs(x-median(x))
tmp=tmp(find(tmp)) % remove zero diffs
mad = median(tmp)

This way, the result on the previous sequence is median([204 6])] whose
result is 99, closer to 116, which permits to detect that one of the
difference is much bigger than the other. 

Thus it make sense, from a statistical point of view, to remove points equal
to the median from the sequence before estimating the mad ?

Regards

Pascal
-- 
View this message in context: 
http://n4.nabble.com/MAD-on-small-data-sets-tp1748291p1748291.html
Sent from the Octave - General mailing list archive at Nabble.com.