Yesterday the work of David Murray, though not his name, was cited in this 
thread. Dr. Murray's name and work have come up before in this forum. I 
can't imagine that the article cited yesterday could be any other than the 
Washington Post commentary piece "Paternity Hype Visits Monticello," which 
appeared on page C1 (the front of the Sunday Outlook section) on Nov. 15, 
1998 -- very soon, as was noted, after the famous DNA report appeared in 
Nature.

Murray is a paternity skeptic who was then directing research at a 
rightward-leaning Washington media-monitoring organization called the 
Statistical Assessment Service. His article can be found in the Washington 
Post online archive, where access requires payment. But a 
copyright-permitted copy appears for free at this URL: 
http://www.tjheritage.org/documents/WP--NatureRush.pdf

Despite what was said yesterday, the article is not a quantitative 
statistical analysis of the conceptions coincidences -- the intriguing data 
showing that TJ was present at Monticello on many but not all of the days 
when Sally Hemings's kids could have been conceived. The article doesn't 
even address those apparent coincidences. Instead, it's a lucid criticism of 
the misreporting that flooded across the media worldwide following 
publication of the Nature DNA report.

"Much of the coverage demonstrated a remarkable flight from careful and 
skeptical reporting," Murray wrote. "All too often, the news stories, 
commentary and analysis transformed an intriguing but inconclusive 
scientific finding into a dead certainty." As far as I know, Murray in this 
article was the first high-visibility whistle blower concerning abuse of the 
special authority of science in the Hemings-TJ controversy.

Murray's article appeared over a year before the William and Mary Quarterly 
published what purported to be a quantitative statistical analysis of the 
conceptions coincidences. That WMQ report confidently claimed the full 
authority of science in declaring that TJ had fathered six Hemings children. 
I started this thread the other day because I'm trying to find out if the 
Thomas Jefferson Foundation is still standing by and citing that report, and 
also because it's possible that others, about whom I don't yet know, have 
attempted similar statistical studies. (And Jurretta Heckscher, maybe you're 
right that I should simply inquire at the Jefferson Library. Thanks for 
suggesting that.)

But Murray's 1998 Washington Post commentary piece deserves to be 
remembered -- even by paternity agnostics and paternity believers if they 
care about the principle of fact-based debate. He understood that many might 
couple the limited DNA molecular findings with other evidence to arrive at 
paternity belief. He respected that, though he didn't buy it. What he 
disrespected, and protested, was the proclaiming of paternity certainty 
based only on limited molecular findings.

When his article appeared, I had penciled myself in as a paternity 
suspecter, if not a paternity believer. Only later -- in part because of 
skepticism-inducing dismay at many journalists' and some historians' 
irresponsibility with scientific evidence -- did I move toward paternity 
agnosticism. But I knew horrendous misreporting when I saw it. Yet during 
the two weeks before Murray's article appeared, I had begun to wonder if I 
was the only person on the planet who actually understood what the Nature 
DNA report did prove and what it didn't prove.

Later Murray was to publish a highly skeptical but mainly nontechnical 
critique of the WMQ statistical report. It was a chapter in an ardently 
paternity-disbelieving anthology with a confrontational subtitle -- The 
Jefferson-Hemings Myth: An American Travesty. There Murray was the first, as 
far as I know, to observe that the WMQ statistical report argues 
circularly -- that it begs the question (in the original but disappearing 
sense of that phrase) with a Bayesian procedure that somehow uses, or claims 
to use, the pro-paternity argument's nonstatistical, qualitative threads to 
validate quantitative statistical simulation results.

But even in that chapter, Murray didn't offer a comprehensive technical 
analysis of the WMQ study. By then I had come to know Murray by e-mail. My 
surmise is that he was daunted not so much by the technical content of the 
WMQ study as by its failure to do what any scientific report is supposed to 
do: provide information about methods so that another scientist can 
replicate the work to check its validity. One of the failings of the WMQ 
study is that it doesn't follow that standard scientific practice. I think 
Murray felt constrained from making certain arguments because he couldn't 
even tell completely what he was supposed to be arguing against. (My friend 
and Jefferson Lab colleague Dave Douglas, a distinguished 
particle-accelerator physicist who makes his living in part by doing 
statistical simulations, has also emphasized this failing in the WMQ study.)

It's worth noting also that Murray did a fine job in a brief debate in the 
Times Literary Supplement with Gore Vidal in late 2000. Vidal had blurred 
the distinction between DNA findings and historical interpretation of DNA 
findings -- a common offense, in Hemings-TJ matters, against the principle 
of fact-based debate. Vidal was haughty and dismissive to Murray, but he 
seems nevertheless to have learned something in the exchange.  In 2003 Vidal 
wrote -- with a simple accuracy that, thankfully, has become more common in 
public discussion of Hemings-TJ -- that DNA testing "establish[ed] 
consanguinity" between the Hemingses and Jefferson.

I was disappointed when David Murray left the Statistical Assessment Service 
and ceased participating in the Hemings-TJ discussion. But I'm glad his name 
and work have come up in this thread. I've lost track of him and don't even 
know if he knows about TJscience.org, which he helped inspire. If anybody 
knows him, please forward this message to him.

Thanks very much.

Steven T. (Steve) Corneliussen
Poquoson, Virginia 

______________________________________
To subscribe, change options, or unsubscribe please see the instructions at
http://listlva.lib.va.us/archives/va-hist.html