Summary of chop-nod analysis telecon of April 22, 2009

Giles, Lero, John, Hiroko

_____________________________________________________________

The math in John's new chi-2 routines.

We reviewed the memo that John recently wrote and posted to the teamsite 
under "useful software" (April 9th).  The question is whether this math is 
a valid way to compute the reduced chi2.  E.g., how does it compare with 
what Mike is doing?

This is complicated stuff and needs more attention from the group.  
Through equation (4) the idea is simple enough: just treat the weighting 
factor "f" as if it is the same as a weighting factor that comes from 
statistical uncertainty.  Beyond equation 4, its a matter of correcting 
for the previously neglected overall scale factor for all the "f" factors.

_____________________________________________________________

Testing John's new chi-2 routines.

John recently posted something on the teamsite that gives results of his 
own chi-2 tests on NGC 1333.  The final result of 1.15 agrees fairly well 
with Mike's map-average values (~2.15) when one considers that Mike crops 
points not observed in all bins.

The hope is to (1) save the time of running Mike's code, and/or (2) 
compute a more detailed "power spectrum" of the systematic errors than 
Mike's "two-point power spectrum" that he did for 1333, and/or (3) 
incorporate automatic outlier rejection.

There are currently two differences between Mike's and John's methods.  
(a) different math (maybe) and (b) Mike chooses bins while John does 
something analogous to a power spectrum.

To move ahead on items 1, 2, and 3 above and also to check the results 
against Mike's, we need the ability to "crop".  Other than that, some 
fancy shuffling around of files should be done, but this does not require 
code modifications.

_______________________________________________________________

M82:

Lero is getting going on M82 again.  Our best advice to him is to follow 
in Mike's footsteps - see Mike's three posted memos.  In particular, does 
the reduced chi2 depend on the number of bins?  If not, we can trust it.  
But note we have to experiment (as Mike did) with different crops.  
Again, refer to Mike's three recent posted memos.

_______________________________________________________________