Archiver > GENEALOGY-DNA > 2009-07 > 1247526993

From: Alan R <>
Subject: Re: [DNA] R1b Origins (was OurEuropeangeographicalblock. . .)
Date: Mon, 13 Jul 2009 23:16:33 +0000 (GMT)
References: <><><><><><001a01ca03e3$7ed888e0$6400a8c0@Ken1>
In-Reply-To: <001a01ca03e3$7ed888e0$6400a8c0@Ken1>

People have mentioned that Tim's approach to markers is useful but means bigger confidence intervals.  What exactly are we talking about in terms of intervals?  Are we talking +/- 1000 years for example or is it better/worse than that?  That sort of level may sound a lot but it is still possible to eliminate certain archaeological events/changes as the smoking gun even if the confidence intervals are fairly big.  For example a date of say 4000BC +/- 1300 years may sound very very broad (5300-2700BC - a 2600 years span) but it is still very useful as only one archaeologically detected potential major population change 'node' (the Neolithic arrival) falls within that period.  So, as long as the confidence intervals are not absolutely huge then the dates may still allow important conclusions to be drawn.  There are essentially only three archaeological 'nodes' of suspected significant continental population movements into the
isles in pre-Iron Age isles prehistory (an 8000 year long  period spanning roughly 8500BC-500BC):

early Mesolithic settlement c. 8500-7800BC
early Neolithic arrival c. 4300-3800BC
early Beaker period c. 2500BC (which comes a very distant third in terms of suspected population input)

Not much continental intrusion beyond a trickle through trade and contact is generally interpreted as happening in between any of these dates (certainly nothing that could possibly explain the dominance of R1b).  So even the blunt instrument of variance dates with fairly big confidence intervals may still pinpoint the smoking gun by capturing one of those three events within its confidence intervals.  However, it is possible that the dates could fall at a point that, combined with big intervals, means more than one of the three archaeological 'events' could be captured. So problems may remain.  What sort of size of confidence intervals are we talking about?  Can some example (say S116 or L21) be quoted with confidence intervals?


This thread: