Concerning this reply message to the Virginia history list, I apologize to those who wish the Hemings-TJ tsunami would finally recede. If that apology doesn't work, I invoke the license that was issued to me in the message appearing further below, and that was reaffirmed later, by Jurretta Heckscher -- to whom I also owe an apology for carelessly misspelling her surname the other day. In 2003, R. B. Bernstein's _Thomas Jefferson_ appeared. Gordon S. Wood, who was said at the time to be at work on the 1789-1815 volume in _The Oxford History of the United States_, called it the best short Jefferson biography ever written. Bernstein echoed other scholars' belief -- including the belief of Jan Lewis -- that solid proof of Hemings-Jefferson parenthood now exists, resting on one nonscientific and two scientific evidentiary "pillars": historical, DNA, and statistical. The statistical pillar is a study of the coincidences between Sally Hemings's conceptions and Thomas Jefferson's sporadic presences at Monticello. My license this morning is to explain my criticism of the very structure of that pillar. The license sensibly stipulates that I be brief and that I use lay language. And indeed I believe I've done that in the essay that I've just posted at the Web site I've just created, http://www.TJscience.org/ . However, although I hope I'm OK on the lay-language requirement, I lack the skill to do the explaining in, say, three hundred words as opposed to the three thousand that I assembled concerning the statistics over the course of several years, off and on. I've been working on the still larger essay itself, in one way or another, for nearly a decade. I started not because I'm interested in the Ultimate Cosmic Answer to the Paternity Question, though I am interested in that, but because I was so offended at _Nature_'s irresponsible handling of the 1998 news of Dr. Foster's molecular findings. Mind you, the offense had to do with the rules of science, not with Hemings-TJ. At the time, I was a tentative paternity believer, not the paternity agnostic that I gradually became later. When the statistical study appeared in 2000, I became an even more serious student of the use and misuse of science in the paternity debate. The title of my newly Web-posted essay is "Sally Hemings, Thomas Jefferson, and the Authority of Science." The stand-first summary says, "Whether or not Hemings and Jefferson had children together, misreported DNA and misused statistics have skewed the paternity debate, discrediting science itself." I wrote the piece originally for the _The American Scholar_, which turned it down. That was some time ago. In 2007, _The New Atlantis_ accepted it, but after they strung me along for many months, keeping me in the dark, and after they finally rewrote the essay in a way that made it unrecognizable, I withdrew it. This week I updated a few things in the essay, created the Web site, and posted it. If you only want to read about the statistical study, please just find the paragraph, a little beyond the halfway point, that begins, "Neither _Science_ nor _Nature_ appears even to have known about the scientific report claiming to erect the statistical pillar, ... ." What you'll find is that I actually tell a story about the evolution of my outlook concerning a scientific paper that appeared sequestered from scientific scrutiny in a humanities journal. The story comes to involve my friends Bill Blackwelder, a biostatistician and a fellow of the American Statistical Association, and Dave Douglas, a fellow of the American Physical Society and a veteran user of the simulation method that underlies the scientific study. Here I'll report that I'm not a scientist. I have two degrees in English. Since 1985 I've worked for and with, and written about, scientists and research engineers in national physics laboratories, universities, and NASA. I also serve as a media advisor to the CEO of a national scientific organization. Today I probably couldn't pass the final exams in the basic statistics courses I took in 1968 and 1983. But Bill and Dave agree with me that the fundamental problems in the Hemings-TJ statistical study don't require formal statistical training to see. About a year and a half ago I published a Sunday commentary piece in the Richmond Times-Dispatch about what I called Hemings-Jefferson science abuse. Annette Gordon-Reed published a letter to the editor criticizing it, but in my view, and in the view of Dave Douglas as expressed in the rebuttal letter that he published, she had missed something crucial concerning the statistical study (and she had missed much else, but let that go for now). She had failed to see that I'm not discounting the conceptions coincidences as important evidence. Of course they're important -- qualitatively. What I was, and am, criticizing is something quantitative, not qualitative. I'm criticizing the false invocation of the special authority of statistical science to claim proof at a level near certainty for what is only a historical interpretation. For all I know, that historical interpretation is right. I'm just a referee blowing the whistle on rules infractions. And again I apologize that I'm so long-winded about it. You've heard that saying, if I had had more time, I could have been briefer? That surely applies to this e-mail message. But I took ten years writing the essay, and it's as short as -- by my lights, such as they are -- I can make it. Thanks very much. I have to go to Washington today, returning late Friday night, but I'll be interested to see if anyone comments. And although I admit in advance that I'll wince with the petty pride of an author, I'll also be grateful for corrections of errors of fact and for criticisms about gray areas. Steve Corneliussen ----- Original Message ----- From: "Jurretta Heckscher" <[log in to unmask]> To: <[log in to unmask]> Sent: Sunday, May 04, 2008 3:20 PM Subject: [VA-HIST] Statistical Science (was Re: Jefferson's Overseer) > On May 4, 2008, at 12:47 PM, S. Corneliussen wrote: >> >> >> True, the misreporting of valid DNA evidence and the outright misuse of >> statistical science originated among people representing science, not >> the history profession, though credulous historians unskeptically >> accepted the statistical stuff. >> >> > Mr. Corneliussen, I assume that your allusion to the use (or misuse) of > statistical science refers to the article by Fraser D. Neiman, Director > of Archaeology at Monticello, that appeared in the William and Mary > Quarterly circa 2000? As I recall, it applied statistical analysis to > the probable dates of SH's conceptions and the known dates of TJ's > presence at Monticello to demonstrate the extreme improbability that > anyone else was the father of her children. > > This did not, of course, absolutely rule out the paternity of some other > man whose presence at Monticello invariably correlated with TJ's. And > Dr. Neiman is of course an archaeologist (and a very good one), not a > statistician. However, along with the DNA analysis, his > statistically-based conclusion is indeed the other piece of > scientific--as opposed to traditionally historical--research that many > historians, myself included, have found compelling. > > I am probably not alone among such historians in lacking the statistical > training to evaluate Dr. Neiman's study as science. If his study is > indeed, in your opinion as a scientist, "outright misuse of statistical > science," could you possibly give us a brief explanation in laymen's > terms of why you believe this to be so? > > If you can, thanks very much. > > --Jurretta Heckscher > > ______________________________________ > To subscribe, change options, or unsubscribe please see the instructions > at > http://listlva.lib.va.us/archives/va-hist.html ______________________________________ To subscribe, change options, or unsubscribe please see the instructions at http://listlva.lib.va.us/archives/va-hist.html