Give and take on "Lancet's Boil"

(Letters exchanged regarding Michael Fumento's article, Lancet's Boil, The American Spectator Online, August 21, 2007. )


It was with some regret that we read Michael Fumento's article "Lancet's Boil" which could not distinguish between a peer reviewed journal article and an uncritiqued presentation made at a huge conference. The paper cited, written by a political scientist, criticizes our study in Iraq for setting aside a neighborhood sample in Falluja where almost a third of residents had died when on average only about 2% of the population had died in the 32 other neighborhoods sampled. In doing so, this lowered our estimate of Iraqi deaths but narrowed the confidence interval. The criticism in this paper argues that the calculation excluding the extremely high Falluja estimate was an attempt by us to prevent having a confidence interval that included zero deaths. The paper argues that outliers could have been that far below our other measures and the exclusion was inappropriate.

First, in spite of Mr. Fumento's article, a -30% death rate is not possible unless you believe that all the dead people from the past 15 years can come back to life. Secondly, imagine that if you took 32 water samples in your child's school and found that arsenic samples were 60% higher than the national standards, and you were 98% sure that you were above the standard. Then you took one more sample that was ten times higher than the standard. The logic endorsed by Michael Fumento is that we do nothing about the water supply because this outlier broadened our statistical confidence interval and we are not sure that we are above the national standard.

Finally, I hope that the Spectator would have more to report on Iraq than trying to discredit a three-year old survey that has since had its findings confirmed by two other surveys.

Mr. Fumento's downplaying of the carnage in Iraq serves no one well.

Les Roberts
Richard Garfield
Columbia University
New York, New York

Michael Fumento replies:

Roberts and Garfield would have us think Kane is a lone wolf. But the Lancet report was also criticized in Science in November 2006, under the title: "Iraqi death estimates called too high; methods faulted." Science is the most prestigious U.S. science publication. They then simply ignores all of my arguments, apparently in the belief that readers will not refer to the original article. Here's a refresher.

We are told that, for whatever reason, it's okay to consider a third of the data a "statistical outlier." That would be like releasing a "national poll" comparing the top Democratic and Republican nominees that excluded respondents in New York, New Jersey, and California. If the findings become useless by including Falluja, then so be it. If researchers can't collect the data they need, they should admit it rather than taking what they do have and pretending it's enough.

They apparently continue to insist, despite the obvious and despite my personal experience interviewing Iraqis, that death certificates aren't necessary because Iraqis don't lie!

Roberts and Garfield ignore my observation that "100,000 deaths during the survey period [in the 2004 report] meant an average of over 180 a day, of which the Lancet attributed more than half to airstrikes. Have you heard anyone ever claim our airstrikes killed over 90 civilians on any one day during the entire course of the war?"

Roberts and Garfield make no effort to explain why he used a pre-war Iraqi mortality rate of 5.0 per 1,000 people annually while, according to the CIA World Factbook, the pre-invasion (2002) rate was over 20 percent higher at 6.07 per 1,000. As I observed, since the study was allegedly looking for excess mortality; therefore the lower the authors set pre-war mortality the higher the excess post-invasion mortality looks.

They also ignore the noted disparity between his data and that collected by the antiwar group Iraq Body Count, which as of August 26, 2007, had a maximum mortality figure of 77,183 – rather below the Lancet's 655,000 figure as of the end of July 2006. In fact, the group issued a stinging rebuke of Roberts's ploy in an October 2006 press release. It noted, among other things, that the Lancet figures are ridiculous because they indicate that, "On average, a thousand Iraqis have been violently killed every single day in the first half of 2006, with less than a tenth of them being noticed by any public surveillance mechanisms..." Moreover, to believe Roberts and Garfield is to believe "Over 7% of the entire adult male population of Iraq has already been killed in violence."

Any one of these is enough to render the Lancet work worthy of nothing more than wrapping fish; I'm just playing pile-on.

Finally, both David Kane and I called for a general release of the dataset. Earlier this year in Nature there was also was a call for releasing the full dataset. Instead, Roberts and colleagues decided only to release them to the "right people," not to critics or potential critics. What is he hiding?

Let's conclude with the end of Iraq Body Count's statement on the Lancet work:

In the light of such extreme and improbable implications, a rational alternative conclusion to be considered is that the authors have drawn conclusions from unrepresentative data. In addition, totals of the magnitude generated by this study are unnecessary to brand the invasion and occupation of Iraq a human and strategic tragedy.


Michael Fumento, a former paratrooper, has been embedded three times in Iraq and once in Afghanistan.

View his Afghanistan photo set. Read Michael Fumento's additional writing on the military, on Iraq, and on the media, and view his Spring 2006 Iraq photos from both the Fallujah area and Ramadi, his Fall 2006 Ramadi photos. View his 2005 Iraq photos.

Michael Fumento has paid for his military trips entirely out of pocket, including roundtrip airfare to Kuwait, hotels in Kuwait, war insurance, and virtually all his gear. Please support him via PayPal Donate or Amazon Honor System via the logos below.


Amazon Honor System Click Here to Give Learn More