Re: systematic errors

On Jul 31, 2009, at 10:05 AM, Michael Nowak wrote:

> Who knew...  I have done things like vary the level of the  
> background during the fits.  (I have actually been using in ISIS my  
> own variation of the XSPEC "corrfile" procedure, except that the  
> normalization constant could be varied during the fit, not  
> iteratively between fits.  Took XSPEC several years more before they  
> changed their functionality to allow that.)  But I've never tried to  
> increase the error bars.  But thinking out loud here, we have,
>
> 	\Delta C^2 = C +  S_d/S_b * B_s + (F_d*C)^2 + (F_b*B_s)^2
>
> where \Delta C^2 is the variance on the data counts, C is the  
> counts, S_d is the data backscale*data exposure, S_b is the  
> background backscale*data exposure, B_s

Sorry, I mean to say S_b is the background backscale * background  
exposure.  I.e., S_d/S_b is the scaling parameter to convert from  
*measured* background counts to the scaled background counts that is  
the relevant estimate of the background for your data set.

-Mike


----
You received this message because you are
subscribed to the isis-users list.
To unsubscribe, send a message to
isis-users-request_at_email.domain.hiddenwith the first line of the message as:
unsubscribe