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

Giles, Lero, Martin, Hiroko

_____________________________________________________________

Lero's analysis of M82 

Giles suggests that we should agree that we when using chi2 we should 
never use input files that have already had their errors inflated by chi2.

Giles realized that much of his previous advice to Lero was based in 
misinterpreting the chi2 maps Lero was sending him.  The chi2 maps that 
are important are the Q and U ones, not the I one.  Giles was looking at 
the chi2 map for I.  The chi2 maps for Q and U are accessed by looking at 
'fits extensions' to the chi2 output file, not the main output map.  
E.g., in SAO ds9 use "open other".
 
In fact, the q and U chi2 maps for M82 look qualitatively similar to what 
was found for IC 384 and L 1527 (for the latter, see 2008 posting on 
outlier rejection) in the sense that there is no apparent correlation with 
flux.  This is different from what was found for 1333, where chi2 was 
higher in the vicinity of 4A and 4B.

_____________________________________________________________