Re[2]: STATISTICAL FREE RELEASE RAD SURVEYING TECHNIQUES.

[sandy_perle@email.fpl.com]

     Jim,
     
     You make a very excellent point. Statistical sampling is only a viable 
     option when one has the process being sampled in control, and by 
     control I mean within statistical control. Is the process stable or 
     not. If it isn't, and there is a high degree of variation in the 
     process, then a random sample is not adequate. To determine if the 
     process is in control, one should have collected enough data to 
     determine the Cp .. or Ck which is the Process Capability. If the Cp 
     or Ck is not within statistical limits, and those limits are within 
     (or less than the administrative limits established) statistical 
     sampling can not be an acceptable tool. In other words, if your upper 
     and lower statistical limits are determined to be .005 units, and your 
     administrative limits are .075, you're already in trouble. To refresh 
     everyone's training regarding statistical control, if the 3 sigma 
     limits are .005, that means that 99.97% of all data will fall within 
     those limits, if in control. Anything outside of that is chance, and 
     does not fall within random sampling. That is the part you were 
     talking about Jim.
     
     Therefore, I too would be interested in hearing more on statistical 
     contamination sampling. Setting a limit for sampling and a limit for 
     acceptability is not statistics ... it's a method of conducting a test 
     only, and whether or not it is valid, depends on a lot of other data 
     and "statistical" tests... with valid conclusions. As stated, I would 
     be interested to hear more, and maybe see more on this subject.
     
     Sandy Perle
     Supervisor Health Physics
     Florida Power and Light Company
     Nuclear Division
     
     (407) 694-4219 office
     (407) 694-3706 fax
     
     sandy_perle@email.fpl.com


______________________________ Reply Separator _________________________________
Subject: Re: STATISTICAL FREE RELEASE RAD SURVEYING TECHNIQUES.
Author:  radsafe@romulus.ehs.uiuc.edu at Internet-Mail
Date:    11/8/95 4:46 PM


At the risk of sounding like one of those folks who hasn't come up to speed yet,
     
I think statistical survey techniques make sense where the deposition of 
contaminant probably occurred as a result of a steady state process over a 
relative wide area.
     
I don't think it is appropriate where the deposition occurred as a result of 
process or behavior that is unpredictable in nature.  Thus random survey of a 
field or warehouse floor may make sense; random survey of a radioisotopes 
laboratory may not be appropriate as researchers have a habit of squirreling 
away 
things for posterity that one would only find with a 100% survey.
     
I would be interested in how one would describe areas where a statistical survey
     
technique is appropriate and areas where it is definitely not appropriate.
     
Jim Barnes, CHP
RSO
Rockwell Aerospace; Rocketdyne Division 
---------------------------------------