Monbiot's article on the Lancet study drew this letter from Gil Elliot:
On the strength of having calculated war deaths around the globe over the past century, I can inform George Monbiot (The media are minimising US and British war crimes in Iraq, November 8) that the Lancet report on Iraqi deaths is deeply flawed. The number of deaths uncovered by the fieldwork, excluding Falluja, was 21; this figure is extrapolated to a population of 20 million to arrive at the calculation of 100,000 deaths. No one who understands the battlefield would accept such a figure based on 21 bodies. Since most of the deaths were from aerial bombardment, any estimate would require a method of sampling and calculating in clusters related to special knowledge of areas of military action and lethal capacity of the firepower. Falluja would have been a valuable cluster in such a study, but the authors left it out of their findings since, by their methodology, the extrapolation from there gave such an incredible result - 200,000 for the Falluja cluster (3% of the population) alone.
Detailed discrepancies are legion. Of the 53 deaths in Falluja, only one is reported as non-violent, ie 2% compared with 76% for the rest of the sample - there is no explanation of this. The number of female deaths in the 100,000 is extrapolated from a base of one or two female bodies out of the 21 - one has to dig deep into the obscurity of the report to discover this. Some 45% of the violent deaths are of children under 13, only 7% of women - hardly believable. Many of the male deaths (46%), the survey confesses, "may have been combatants", although this is a survey of civilian deaths.
Roberts and Garfield correct Elliot's errors in a reply
Gil Elliot's critique of our Iraq casualty estimation (Letters, November 10) raises many excellent points but contained several errors. Most importantly, the line "The number of deaths uncovered by the fieldwork, excluding Falluja, was 21" is simply wrong - that number was 89. Second, our random sample of 988 households in 33 neighbourhoods represents the entire population and has no inherent errors when estimating deaths. The fact that aerial bombing, which causes deaths in clusters, was a significant cause of death contributes to the imprecision of our findings, but does not necessarily make the estimate high or low.
As hinted by Elliot, we strongly suspect that our 100,000 estimate is low. While deaths reported were confirmed with death certificates more than 80% of the time, families may have hidden deaths. The shame of burying wives and mothers without ritual may explain the lack of adult women. Our study has many limitations. The occupiers can and should improve upon our efforts to acknowledge and respect those lives lost.
I have a suggestion. A new study!
Off you go, Slick. I hope you find the natives friendly.
The authors of the Lancet study also suggests a new, better study. In fact, the say so in the letter above:
"The occupiers can and should improve upon our efforts to acknowledge and respect those lives lost."
It's just that the occupiers discourage any serious attempts to find out how many people thay have killed.
If the UK/ USA are so sure that the Lancet study is wrong, why not do their own? Then if it shows teh Lancet to be wrong, then they have a brilliant PR coup on their hands.
Of course, if it confirms it, then they dont. But since they are so sure they are right in the first place, why not do the replica study to prove it?
guthrie,
You seriously think any study done by the Americans will be taken seriously by the rest of the world? It has to be done by an independent party, which is what the Lancet study authors asked for at the end of their study.
As for Roberts answer to this letter, I have to take issue with this:
"Second, our random sample of 988 households in 33 neighbourhoods represents the entire population and has no inherent errors when estimating deaths."
No inherent errors? Balogna.
No inherent errors = unbiased.
They're right on that point.
If "the" Americans are the President's discredited spinmeisters then of course they can't expect to be taken seriously. But the involvement in the Lancet study of well-qualified Americans, from the Center for International Emergency Disaster and Refugee Studies, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA and the School of Nursing, Columbia University, New York, NY, USA, did no harm at all to the study's credibility.
*[Deleted. Seixon, you have your own thread where you can post rubbish about statistics. Please use it. Tim]*
Kevin,
Clearly I was talking about the "Americans" such as what was suggested, that the American military or American government conduct a study. Don't let English comprehension get you down!
It's not clear to me that guthrie had a government study in mind. An independent study, even if commissioned by government, would obviously carry more weight.