NavList:

RE :
http://www.fer3.com/arc/m2.aspx?i=115812
[NavList] Re: Friendly challenge : Jupiter Lunar Exercise 09 Feb 2011
From: m_burkes---com
Date: 8 Mar 2011 07:19

Hello Mike,

Nice that you enjoyed the various contributions to this "unorthodox" Lunar Exercise. And thanks again to all Contributors.

As regards the option to "reject" outfliers, you might remember our recent long and interesting thread/debate on that subject a few months ago.

There were a few different view points here, with - in particular - one of them recommending to reject observation(s) only and ONLY IF totally out of reasonable range.

Any other observation should be retained since - even in its apparent "inaccuracy" - it very importantly indicates some kind of "just fair" environmental observing conditions.

In other words, rejecting one obervation which is not (exactly) in line with the others is generally not the best approach since the remaining part of the observations set induces a TOTALLY FALSE sense of "better" quality.

For this reason, and according to the view point hereabove, plain data averaging (of all observations) without having to first graph each of them individually is certainly a quite valid "data management" approach. Plain data averaging also has the advantage that it totally eliminates the need for any "slope" adjustment since the averaged values always belong to any data fitting least square straight line, whatever its slope, whether observed or forced.

Best Regards

Kermit

Antoine M. "Kermit" Couëtte

----------------------------------------------------------------
NavList message boards and member settings: www.fer3.com/NavList
Members may optionally receive posts by email.
To cancel email delivery, send a message to NoMail[at]fer3.com
----------------------------------------------------------------