Friday, January 4, 2008
Experts are still combing over the data and methodology behind that infamous and widely criticized Lancet report on casualties in Iraq. A new article has been published by National Journal: Data Bomb. Michael Rubin posts one of the authors' summaries at The Corner:
George Soros funded the survey. The U.S. authors played no role in data-collection, and did not apply standard anti-fraud measures. The chief Iraqi data-collector had earlier produced medical articles to help Saddam’s anti-sanctions campaign in the 1990s, and said Allah guided the prior 2004 Lancet/Johns Hopkins death-survey. Some of the field surveyors were employed by Moqtada Sadr’s Ministry of Health. The Iraqis’ numbers contain evidence of fakery, and the Lancet did not check for fakery.
Dean Esmay has posted the YouTube of Lancet editor Dr. Richard Horton on a political rant: Indicting The Lancet on Iraq
Without having looked at the original article itself a few things that come to mind are:
1) In most scientific journals the author(s) have to disclose any financial or corporate ties they may have.
2) Often, they must also cite the source of their funding for the research (this one is a bit less common)
3) Is The Lancet peer-reviewed?
It is peer-reviewed, though, according to the article, the process was "rushed" in order to make it to publication in time for the presidential election.
jaws:
Another point, it seems to me, is that a reputable study should be willing to make its raw data available to outsiders. (That way people can see if the data collection methods make sense, if statistical errors were made in reaching the conclusions, and so forth.)
From what I read, the Lancet folks have consistently refused to share their raw data. That ought to make anyone suspicious.
respectfully,
Daniel in Brookline