RE: 3X background decision level

["Fulmer, Philip" <philip.fulmer@cplc.com>]

Will,

I've never seen it written in a formal report or textbook, but I have
seen it written in as a license condition in an agreement state
(although I've always seen it as twice the background, not three times
the background).  The logic is that the background follows a statistical
distribution which is centered about the mean value.  If the
distribution is roughly symmetrical, the distribution from 0 to the mean
should be almost a mirror image of the distribution from the mean to
twice the background.  It's not exact science, but if you set the
decision level at twice the background, this is almost the same location
as the classic critical level (Lc) that you would calculate using the
standard formulas, which is the threshold for declaring that activity is
present.  

I would hesitate to use three times the background because in that case,
you're approaching the MDA and using a decision level at that point
would cause a lot of false negatives (although you would almost never
get a false positive!!) by having a decision level set too high.

Of course, the assumption in this is that the value you use for
background is fairly close to the true mean and not too far on the upper
side or lower side of the distribution.

Philip

__________________________
Philip C. Fulmer, PhD, CHP 
Carolina Power & Light Company
Harris Energy and Environmental Center
3932 New Hill-Holleman Road
P. O. Box 327
New Hill, NC 27562-0327
philip.fulmer@cplc.com
(919) 362-3363      

>----------
>From: 	William McCabe[SMTP:WMCCABE@tnrcc.state.tx.us]
>Sent: 	Thursday, July 30, 1998 2:09 PM
>To: 	Multiple recipients of list
>Subject: 	3X background decision level
>
>Can anyone provide me with the details/basis for the thumbrule or
>practice of using 3 times (or sometimes 2 times) background as a
>detection (or decision) limit for portable survey instruments?
>
>Thanks in advance.
>
>- Will McCabe
>
>
>
************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html