(Please note: This is not a re-hash of did-they-or-didn't-they. It's a
historiographical discussion -- or maybe a techno-historiographical
discussion -- that still matters and that can still lead to useful clarity
about what we know and what we do not know. It also presumes some
familiarity with the Fraser Neiman statistical study.)
Here are a few comments about Mr. Dixon's responses concerning the Neiman
statistical study:
* Mr. Dixon charges that the "William & Mary Quarterly published [the Neiman
statistical] article without peer review," but ca. 2000 the WMQ editor
assured me with great confidence over the phone that it had indeed been
peer-reviewed. For peer-review confidentiality reasons he could say no more.
I note that a Neiman footnote acknowledges two distinguished demographers.
* I'm not sure what Mr. Dixon means when he says "I am not aware of any one
who has followed Fraser's methodology to determine exactly what it was he
did." It's true that among the study's scientific failings is that it
doesn't document its methods so that others can test it by seeking to
replicate it, as the physicist David R. Douglas has noted. It's true that
that caused puzzlement for Dave and for the biostatistician William
Blackwelder, the other scientist on whom I called for what I wrote at
TJscience.org. But that methods-documentation omission doesn't make the
study or its methodology immune from criticism or analysis -- of which the
study and its methodology got plenty from Dave and Bill, and from me. I
apologize for this pompous repetition about their qualifications, but these
two scientists are fellows of the American Physical Society and the American
Statistical Association, respectively. In science, such recognitions are
substantial. It is simply not true that no one has sought to follow the
Neiman methodology, if that's what Mr. Dixon is saying.
* The Neiman study involved not only Monte Carlo simulations done by
computer -- as Mr. Kukla framed it and as Mr. Dixon repeated -- but also the
application of Bayes's theorem. That's a method of modifying a statistical
result to account for other information -- in this case, the pre-existing,
nonstatistical threads of the pro-paternity argument. The Bayesian approach
is sometimes controversial. To my knowledge David Murray, whose name came up
the other day in this forum, was the first to argue that with Dr. Neiman's
undocumented Bayesian procedure purporting to validate the simulation
results, Neiman simply argued circularly -- that is, he begged the question
(in the original sense of that phrase, not the new sense of "pointed out a
question to be asked"). Dave and Bill agree -- and they are confident that
they'd still agree even if Dr. Neiman had revealed how he actually applied
Bayes's theorem.
* The important information in Mr. Dixon's final two sentences relates
closely to what I'm saying about the study's Bayesian dimension. Mr. Dixon
wrote: "It is interesting to note that Fraser commented in his study that if
one believes the paternity hypothesis is 'false on other grounds,' then his
study would be of no consequence. I guess the obverse of that is if you
first believe the paternity story, then you can rely on Fraser's study."
* As to the skimming and the qualitative appreciation of the raw data as
opposed to the quantitative statistical science: I have seen Cinder Stanton
make the point that whatever might need to be said formally within the realm
of statistics, the raw data on the conceptions coincidences are intriguingly
suggestive. (That's not a direct quotation.) I agree with her. But in an
important way that doesn't matter any more. It's the status quo ante for the
history profession now that leading historians have backed a quantitative
scientific claim about the data. That is, we knew about the qualitative
dimension from Winthrop Jordan long ago. Nowadays the transformed claim is
that Lofty Science Itself shows that these data add up to outright proof of
paternity. In my view that's a strong claim that historians -- especially
those whose eyes have glazed over -- should either understand and support or
leave unmentioned.
* I repeat that not only do we have the Bernstein reliance on the Neiman
study, but the reliance also of Jan Lewis, whose praise of it I quoted the
other day from her introduction to that 2000 WMQ collection of post-DNA
historical essays by Joseph Ellis and Annette Gordon-Reed and others.
Some of what I've said here I was repeating. That might be wrong of me, but
it was not an accident. One thing I have not said in a while in this forum
also needs repeating: yes, it's true, I'm not a scientist myself. Some who
believe strongly in the authority of science might find that a drawback.
Maybe it is. But anyone who says so is probably also validating the most
fundamental claim I'm offering: that the authority of science is special and
that it matters intrinsically. (Again: It matters intrinsically here whether
or not SH and TJ were parents together.)
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
|