Date: Tue, 13 Jun 2000 10:52:48 -0500
Reply-To: "Lambert, Bob" <Bob_Lambert@AFCC.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: "Lambert, Bob" <Bob_Lambert@AFCC.COM>
Subject: Re: Convincing non-believers
Content-Type: text/plain; charset=us-ascii
Is not the variation within the lot the driving force? So a larger
variation will demand more samples than a smaller variation. Thus, back to
the question of homogeneity. For example, if it is known that the process
that produced the lots of uranium oxide maintains a "small" variation (e.g.
- from in-process sampling), then that would provide the reasoning that the
reader was asking about.
Salt Lake City, Utah
> -----Original Message-----
> From: Tom Slack [SMTP:tom_sas@DUESOUTH.NET]
> Sent: Tuesday, June 13, 2000 10:17 AM
> To: SAS-L@LISTSERV.UGA.EDU
> Subject: Re: Convincing non-believers
> Didn't want to get into the "fine print" of sampling, and especially the
> details of this project. Let's just say the heart randomizes blood and
> forklifts randomizes barrels of depleted uranium oxide.
> Tom Slack
> Aiken, SC
> At 11:19 AM 6/12/2000 -0600, you wrote:
> >Blood is relatively homogeneous; if you draw one sample you're going to
> >get the same constituents as if you draw 60 samples.
> >What reason was there to think that the uranium was also homogeneous?
> >Were you melting it and mixing it all up together?
> > >>> "Tom Slack" <tom_sas@DUESOUTH.NET> 06/12/2000 8:58 AM >>>
> >I had a similar experience at a DOE facility in South Carolina. Everyone
> >involved kept asking the same question: "What percent of the Lot do we
> >test?" I kept saying that the number of samples does not depend on the
> >Lot size.
> >We had a meeting where I was greatly out numbered by nonbelievers. One
> >woman was going to have a baby soon and one man was tall and robust. I
> >asked the mother to be if when she had her baby, was it likely the baby
> >would need a blood test. She said yes. Then I asked the man if, for
> >similar blood test, would he need to give about 60 times as much blood.
> >turned a little pale and said no. They approved my sampling plan and we
> >were able to have drums of depleted Uranium declared as nonhazardous
> >In the short run, examples seem to work. In the long run, the American
> >Society for Quality has a Power Point presentation called "Statistical
> >Thinking" which treats Statistics as a philosophy and is well accepted
> >most people. Let me know if you would like the slides.
> >Good Luck,
> >Tom Slack
> >Aiken, SC
> >At 08:19 PM 6/9/2000 +0000, you wrote:
> > >I am involved in a study where members of the public that do not trust
> > >the government agency responsible, The Deparment of Energy.
> > >
> > >Once the study is complete there will be a need to convience the public
> > >the lack of or extent of hazard to human health.
> > >
> > >How do you explain the results from sampling a large area that are
> > >poisonous in extremely small quantities, long lived (10s of thousands
> > >years), and extremely difficult to detect (you can't smell it, see it,
> > >touch it and the chemistry is expensive)
> > >
> > >How do you convince them without teaching a class in statistics?
> > >--
> > > William S. Kossack
> > > Westminster, Colorado
> > > firstname.lastname@example.org