GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-03 > 1110417637


From: (John Chandler)
Subject: Re: [DNA] reliability of the DNAPrint test
Date: Wed, 9 Mar 2005 20:20:37 -0500 (EST)
References: <410-22005339221619578@earthlink.net>
In-Reply-To: <410-22005339221619578@earthlink.net> (ernie5823@earthlink.net)


Ernie wrote:
> I'd really like to know the answer to this, also - explained in terms a
> person with no statistics background can comprehend. My early on 2.0 test,
> #0145, shows 0% NA & 0%SA, but the "rings" go further away from the dot on
> the NA line than the SA line.

Think of the confidence rings as an archery target with a bull's eye
in the center, but with a special twist -- the target isn't circular,
and it's tilted to one side, and any part of the target that falls
off the edge of the triangle is simply chopped off. If you look at
really "well-mixed" folks' test results, you can see this obvious
pattern because little or nothing is chopped off. In your case,
however, most of the target is missing, and you have to stretch your
imagination to see the pattern. (One other complication: if the
red dot falls off the edge of the triangle, they pick it up and
stick it down on the nearby edge, and they don't tell you that
they have done this.) One reason why the rings go further along the
NA line than along the SA line, and perhaps the only reason, is the
tilt of the target that I mentioned. To put it another way, the test
is inherently more uncertain in detecting NA than SA.

John Chandler


This thread: