GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2003-11 > 1069474927


From: "Peter A. Kincaid" <>
Subject: Re: [DNA] Re: STATISTICAL DISCUSSION at DNA-D Digest V03 #709
Date: Sat, 22 Nov 2003 00:22:07 -0400
In-Reply-To: <1ad.1cac01e5.2cf02c76@aol.com>


At 10:05 PM 21/11/03 EST, you wrote:
>Referring to the professional-level discussion involving > Bayes' theorem
for
>> conditional probability:
>>
>> Pr(2M | 2k, t) = Pr(2k | 2M, t) * Pr(2M | t) / Prob(2k | t)
>>
>The discussion brought to mind a question I have often entertained while
>reading DNA-D Digest: I really wonder how many of those who engage in the
>colloquium have any feel at all for the laws governing statistical
inference. I am one
>of those who are really observing from the outside. It seems much too
complex
>a subject for me to wrap my brain around. I have a book in my library,
>titled: "How to Lie with Statistics" and in that volume is this aphorism:
"There are
>three kinds of lies: lies, damned lies and statistics."


One has to remember that the whole system is based on assumptions; namely
axioms. Change you assumptions and you get a whole new perspective. One
just
has to walk down the road of non-ecludian geometry to get my drift. This
is where I bailed out in university. Made me wonder whether it was all
invented to waste a lot of time avoiding good old fashioned common sense! ;-)





Peter A. Kincaid
Hampton, NB, Canada


This thread: