RE: Control Charts for Alpha Counters

[Christoph_Hofmeyr/CNS1@cns.co.za]

 Radsafers, Jay MacLellan,
Pardon the belated reaction due to absence: Yes, absolutely. I did not want
to convey that one should invest less in determining background. On the
contrary, I wanted to point out that one might feel inadequately rewarded
for the effort to determine the mean background accurately.  It is crucial
to get a good average value, a "good estimate of the mean", based on as
many events (counts) as one can afford.   In this way, the uncertainty on
the background variance is reduced, but the magnitude of the background
variance, which is relevant to the sample run, is determined by the mean
background rate times the "sample counting time".  Expressed as a
coefficient of variation, this is much larger than for the total background
determination.  This could give one the feeling that you did not get full
value for your backgound measuring effort.   But, as a consolation of
sorts, variances add absolutely when subtracting background.   If many
background runs are recorded, each for the intended "sample counting time",
the results can in principle be used to verify the Poisson distribution
pertaining to the background for this selected sampling time.
Disclaimer: My personal opinion.

Chris Hofmeyr
CNS
Centurion
South Africa
chofmeyr@cns.co.za


You wrote:


Subject:  RE: Control Charts for Alpha Counters


>Chris Hofmeyr wrote "If one can be confident that the background is also
Poisson
>distributed, it is a big advantage.  The beauty of a Poisson distribution
is of
>course that the average is equal to the variance..."

I agree with everything you say.  I just want to emphasize that the true
variance equals the true mean, but you don't have a good estimate of the
variance without a good estimate of the mean.  If you establish a control
chart
based on the variance of an inadequate sample population, you may be
setting
yourself up for trouble.

Just my opinion.

Jay A. MacLellan, CHP
Radiation and Health Technology
Pacific Northwest National Laboratory
Phone:  509-376-7247
Fax:      509-376-2906
Email:   jay.a.maclellan@pnl.gov

************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html







************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html