Archiver > GENEALOGY-DNA > 2009-07 > 1247873015

From: Alan R <>
Subject: Re: [DNA] For Age Estimates beware of tossing STRs away
Date: Fri, 17 Jul 2009 23:23:35 +0000 (GMT)

I agree with that and, as I posted earlier, the fact that your ditching of medium STRs has consistently made the TMRCA dates older suggests they really are.  Another thing I was thinking is if the central dates for clades previously calculated to have similar ages (S116, L21, S28 etc) keep consistently coming in sharing very similar (but older) central ages in the new calculations, does this give us any comfort that the central ages may be close to reality.  Essentially if repeated separate calculations keep coming up with the same central ages despite the wide confidence intervals then do the central ages gain more credibility and does this reduce the need to consider the full confidence interval range.  Also if confidence intervals are crucial then why has there been a strong tendency to simply quote a single central date and no intervals in DNA dating?  One example is Oppenheimer who quoted absolute dates and interpreted them as though
confidence intervals didn't exist.  The intervals were actually quoted in the notes at the rear of his book and often amounted to +/- thousands of years that if he had taken into account would have meant his very strident statements would have been impossible to make.
Tim said,
In general, I would rather have wider confidence intervals and believe that my TMRCA estimates are closer to the
truth than have TMRCA estimates that are obviously incorrect and yet have narrow confidence intervals.

This thread: