29218: Kondrat (comment) re: 29203: De Verteuil (and Durban) on the science of the Lancet Study (fwd)

[Bob Corbett <corbetre@webster.edu>]

from Peter Kondrat   kondr8@gmail.com:


Contrary to Patrick de Verteuil's claim, I haven't seen anyone on Corbett
insisting that the Lancet study data are "unassailable gospel truth,"
although we have had Lance Durban proclaiming his belief that the numbers
were simply made up and then published in one of the most respected
peer-reviewed scientific journals in the world. His is an extreme view, less
an opinion than a wild hunch.  That's fine; I know people who think the moon
landings were fabricated in a TV studio.  Some who harbor unconventional
beliefs entertain us with theories explaining how they come to their
conclusions; but Mr. Durban just seems to fold his arms, say he doesn't
believe them durn numbers, and let it go at that.

I am not so allergic to data. While I won't accept the Lancet fiindings as
gospel, I've seen no evidence that would lead me to doubt their veracity.
Overall, scientific studies like this one are healthy for Haiti's social and
political future. As Haiti moves away from its dysfunctional political past
and toward a working democracy, it will also distance itself from the
purveyors of opinion, rumor, gossip and ideology disguised as "information."
The Lancet study is a step in the direction of clarity and transparency.
It's clearly a threat to those who lurk in that other, murky world.

Which leads us to Mr. de Verteuil and his "analysis."  I've actually read
the ten-page Lancet study, twice now. It does include a frank discussion of
the limitations of the sampling method used. But it did *not* ask survey
respondents about what happened to their "relatives," as Mr. de Verteuil
claims it did. The survey gathered data about "household members," and the
report discusses at length the pitfalls and shortcomings of using that
concept in Haiti. There is plenty that could be commented on in responding
to or criticizing the study, but it would seem that blind criticisms of the
study, that ignore what the study actually says, cannot be considered
seriously.

Likewise, Mr. de Verteuil's criticism that "surveys taken in Haiti,
particularly with Haitians of rural extraction, are inevitably flawed by the
rural Haitian tendency to tell the questioner what they believe he wants to
hear" can be made of any survey, in any country. (We could just as easily
distill his criticism to "surveys . . . are inevitably flawed by the . . .
tendency to tell the questioner what . . . he wants to hear.") His personal
anecdote is interesting, but hardly persuasive; I think it may be human
nature to put on a good face for an inquiring stranger. And Mr. de
Verteuil's "insight" about credibility among survey respondents is old news
indeed to anyone with even a little knowledge of the debate about the
reliability of surveying as a data-gathering technique.

Patrick de Verteuil is partly right. There do seem to be two types of
Corbett Listers commenting on the Lancet study: those who have read it, and
those who have not. Alas, he appears to fall into the latter category.

Am I the only one who thinks it fair to ask Corbett Listers to actually read
a study they purport to criticize?