[Ifeffit] choosing k-weight

[SCalvin100 at aol.com]

Peter,

Of course, the bottom line is that the k-weight shouldn't affect the values 
of the parameters Ifeffit finds. If it does, it's a sign that the parameters 
should not be trusted. Often, this happens when the k-range chosen to fit is 
inappropriate (extends up into a range where the data is dominated by noise, down 
into a range where the background is poorly chosen, or across a small edge 
due to a contaminant, for example). I always test my "final" fits at all three 
k-weights separately to make sure the values of the parameters stay fairly 
stable.

--Scott Calvin
Sarah Lawrence College
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20040220/bf87091f/attachment.html>