100,000 dead?

November 5, 2004 8:20 PM

The Lancet recently claimed that up to 100,000 Iraqi civilians are dead as a result of the war. Believable or not?
posted by homunculus (29 comments total)

Yet another link: Consider just this: The sample size was so small that the range for deaths was a humongous 8,000 to 194,000. So Roberts and friends just split the difference.

From Zmag: Thus, it is 95% likely that the true rate of excess deaths is between 8,000 and 198,000.

Sounds like guessing.
posted by Ayn Marx at 8:51 PM on November 5, 2004

Oh yes, Ayn, but that 100k number is soooo much cooler to use for propaganda. Just like the authors intended with the timing of their report.

You know, the political impact of this study would've been even more profound for the anti-war argument, if simply the 8000 number was used. Even that number is disgusting, and then revealing that it could just as likely to be 100,000 drives the point home.

I also see no attempt to quantify the "excess deaths" resulting from the use of civilians as human shields, especially because that argument is used by the other side of the rhetorical chasm.
posted by reality at 9:21 PM on November 5, 2004

I'd pretty much dismissed it after Kaplan's piece, but the Zmag article has me wondering. Still, I think Iraq Body Count is more reliable.

Meanwhile, it looks like the assault on Falluja is imminent.
posted by homunculus at 10:05 PM on November 5, 2004

Sounds like you don't understand the absurdity.

That's like me surveying one Iraqi household who happened to have a family member die, and then going on and extrapolate that I am 95% confident the actual death rate is somewhere between 1 and 10 million. Of course, I would then publish the middle number (5 million) in my favorite leftwing publications.
posted by VeGiTo at 10:07 PM on November 5, 2004

Ooo, I can hardly wait to watch the vilification of the Lancet as an evil ultra-liberal propaganda mouthpiece!

Hey, maybe they'll decide all medical science is evil (tools of the Darwinists, don't you know) and die of typhoid or ingrown hairs something?
posted by hattifattener at 10:09 PM on November 5, 2004

TNR interview with Gilbert Burnham.
posted by homunculus at 10:48 PM on November 5, 2004

Of course it's believable. That's only 2 football stadiums full of people - small fry by 20th century war standards.
posted by crunchburger at 10:58 PM on November 5, 2004

I'm no statistical expert but it sounds like the cluster sampling was making the best attempt at getting a handle on the number of deaths within the warzone. I've read that in Darfur scientists had to calculate the death toll using a census. It sounds like trying to track this kind of information is inherently difficult and variable.

From the Slate article:

The IBC estimates that between 14,181 and 16,312 Iraqi civilians have died as a result of the war—about half of them since the battlefield phase of the war ended last May. The group also notes that these figures are probably on the low side, since some deaths must have taken place outside the media's purview.

So, let's call it 15,000 or—allowing for deaths that the press didn't report—20,000 or 25,000, maybe 30,000 Iraqi civilians killed in a pre-emptive war waged (according to the latest rationale) on their behalf. That's a number more solidly rooted in reality than the Hopkins figure—and, given that fact, no less shocking.

It would be interesting if we could get a death toll from the Military, but it is impossible since they don't track the number of civilians killed. It's understandable why they don't do this - they probably fear the number would be used as propaganda by the insurgency. But as it is, the absence of a number will also serve as fuel for propaganda.
posted by rks404 at 11:19 PM on November 5, 2004

The 100,000 estimate is entirely believable, and those who say that the range is somewhere between 8,000 and 200,000 just don't understand the nature of medical statistics. Those who appreciate the work that goes into doing these reports love what these people have done. It is exhaustive and groundbreaking in their field.

While one of the authors say that there is a statistical possibility, however slight, for casualties to be much lower, he also says the following:

"violence accounted for only 2% of deaths before the war and was the main cause of death after the invasion. That is something new, consistent with the dramatic rise in mortality and reduces the likelihood that the true number was at the lower end of the confidence range. Secondly,
there is the Falluja data, which imply that there are pockets of Anbar, or other communities like Falluja, experiencing intense conflict, that have far more deaths than the rest of the country. We set that aside these data in statistical analysis because the result in this cluster was such an outlier, but it tells us that the true death toll is far more likely to be on the high-side of our point estimate than on the low side."

Also, keep in mind that these deaths are not all caused by the US military, but by weapons discharged by the insurrection as well, or kids playing with loose ordinance. Indeed, violence has become the most common cause of death in Iraq. To think that approximately 1300 Coalition soldiers have died in Iraq, with approximately 9000 WIA without many, many times that number of dead Iraqis is naive. A report by the Iraqi Ministry of Health has already said in their very small sampling of deaths in Baghdad hospitals that well over 2/3rds of the deaths were caused by Coalition forces as opposed to insurgent violence. By that estimate, we're probably looking at about 70,000 dead Iraqis by Coalition hands, which is entirely consistant with what I know about the war from the soldiers I know over in Iraq.

I have a friend who works over in Iraq for a company that hires about 50 Iraqis. One of them died in a carbombing, and another was almost killed by US indiscriminate fire. Another was kidnapped, but managed to escape from his captors... so yes, Iraq is horribly dangerous and violent, and the Coalition presence contributes greatly to make it so.

Plug that into your statistics and extrapolate it.
posted by insomnia_lj at 12:23 AM on November 6, 2004

Here is the link for the report indicating the coalition causing more Iraqi fatalities than the insurgents.

I should say that "very small sampling of deaths" is actually about 3500 Iraqi deaths, where they knew the name of the deceased and how he died, and where the doctors actually had the time to write down the information. In many cases, records stopped being kept after the invasion.
posted by insomnia_lj at 12:32 AM on November 6, 2004

Another thing that should be noted for those who think that Iraq Body Count is somehow more reliable...

They get the vast majority of their figures from English-language news reports.

So, the question becomes this: what percentage of deaths in Iraq are reported by the press?

In order for the findings published in the Lancet to be relatively accurate, the western press must only report approximately 20% of the Iraqi deaths that occur in Iraq. And you know what... I think this is entirely probable. When was the last time you heard of Iraqi children dying after playing with unexploded ordinance, people being killed at checkpoints, or heard about *ALL* the miscellaneous one-off Iraqi deaths throughout the country? They're simply not newsworthy by Western standards, but they all still count.

Frankly, if our media is only missing 20% of the deaths, I would be very much surprised.

So, while you can say that IBC is more verifiable, it is highly unlikely for it to be more accurate.
posted by insomnia_lj at 12:52 AM on November 6, 2004

It also helps to apply a little common sense.

If IBC has identified 14K to 16K probable combat casualties, it's logical to assume that "excess deaths" due to other factors would tack a few deaths onto that. Those other factors include things like degraded infrastructure (including devastated medical system and absent medical supplies), radically increased crime rate, radically increased rate of plain old violence, and under-reported casualties.

So anyone seriously suggesting that the actual figure has any real probability of being 8K is being pretty profoundly disingenuous.
posted by lodurr at 1:51 AM on November 6, 2004

Excellent post. I'd wondered about that survey for a while: seemed a little high.

The Lancet also screwed up recently over the MMR controversy.

I think they're in a awkward position, being somewhere between a scientific journal and a popular magazine. They really ought to stop this business of rushing out studies without proper peer review so they can break the story. The Iraq thing isn't so important, but the MMR scare caused huge numbers of people to avoid vaccinations. A little more responsibility please...
posted by TheophileEscargot at 3:48 AM on November 6, 2004

You know, the political impact of this study would've been even more profound for the anti-war argument, if simply the 8000 number was used. Even that number is disgusting, and then revealing that it could just as likely to be 100,000 drives the point home.

This is an entirely inaccurate statement and unfortunately I think a lot of people will interpret the data this way. A confidence interval doesn't suggest an equal probability of the actual number being anywhere within the interval. The actual number of increased deaths is not equally likely to be 8,000 as it is to be 100,000. An accurate statement would be that the actual number is equally likely to be 8,000 as it is to be 194,000. It is much more likely to be 100,000 than either of those numbers.

I don't have the grasp of the full statistical process used to produce these numbers, but in very rough terms, a 95% chance that the number is between 8,000 and 194,000 means there would be a 90% chance that the number is between something like (best attempt at educated guesswork:) 20,000 and 175,000.

(Note: those numbers aren't pulled from nowhere. I approximated the ratio of the margin of error in a 90% confidence interval to a 95% CI at about .85 based on this data and then applied it to the death numbers. I'd love to see someone who actually knows what they're doing produce some more accurate numbers on a 90% CI.)
posted by Wingy at 3:49 AM on November 6, 2004

Daniel at Crooked Timber, pretty much demolishes the existing criticisms of the report (and hosts a rether civil and to the point discussion on the topic as well).

Let me also point out that from what I've seen those that criticise the statistics are people with no background in medical statistical research at all (I'd be interested to see an exception...)

From what I've seen, the study takes pains to err on the low side. I wonder if similar methods weren't used to determine the death toll in similarly difficult circumstances...
posted by talos at 4:26 AM on November 6, 2004

TheophileEscargot, the Lancet remains a very prestgious publication. Indeed a retraction of a previous study, is both science at its best and fairly common in such situations.

The popular misconception that scientific journals don't publish erroneous theories or eventually disproved theories, is grounded on a serious misconception about the nature of the scientific method.

There is nothing to suggest here that the peer review process was anything but proper. Indeed there is nothing to suggest that the methodology was anything but standard.
posted by talos at 4:35 AM on November 6, 2004

Of course, I would then publish the middle number (5 million) in my favorite leftwing publications.

funny how right-wingers aren't this flippant when discussing white people murdered in Lower Manhattan three years ago. 3,000 dead! never forget!

bah.
some of us have equal compassion for all dead. NJ firemen or Iraqi shopkeepers, their deaths are equally tragic for some of us. and equally appalling.

Sounds like guessing.

it does. those damn "Iraquans" have been shattered in pieces so tiny by the freedom-loving American bombs that one cannot guess precisely how many corpses we're talking about here. it's just, you know, a lot. but then, they're all dark-skinned, those corpses (well, they were before they got all charred, but I'm digressing)
posted by matteo at 4:43 AM on November 6, 2004

Oh yes, Ayn, but that 100k number is soooo much cooler to use for propaganda. Just like the authors intended with the timing of their report.

Considering the count of found WMDs remains at 0, the failure of this war was solidified once the body count hit 1.
posted by XQUZYPHYR at 5:40 AM on November 6, 2004

Tommy "We don't do body counts" Franks notwithstanding, if the US military is obliged to fight an insurgency long enough, the body count will be reinstituted as a means of providing tangible results.
posted by alumshubby at 6:33 AM on November 6, 2004

funny how right-wingers aren't this flippant when discussing white people murdered in Lower Manhattan three years ago. 3,000 dead! never forget!

Yes, every one of those people was white. Really. Quite remarkable, but there you have it.


And remember:

Criticism of a methodology == flippancy about the death of brown people.
posted by Ayn Marx at 7:47 AM on November 6, 2004

Also remember:

Picking one word from a sentence and extrapolating a criticism from it == Avoiding the issue.

Even if 10%, 50% or 100% of those who died on 9/11 where black, white, yellow or tartan - 3,000 dead Americans are not worth 100,000 dead Iraqis, 8,000 dead Iraqis or 3,001. It's not even worth 1*.

You should already know this.

*unless it was Saddam or his worthless sons.
posted by longbaugh at 8:04 AM on November 6, 2004

so perhaps matteo didn't express his point quite as eloquently as he would have wished, but you should still get what he was trying to say.

in other words, there was no cold and clinical discussion of how many people might have actually died on 9/11, nor was there discourse on the efficacy of the method used to determine the number of deaths.

but when it comes to the iraqis, here we are just throwing out numbers as if each digit didn't represent a life that has been ended.

in all their "prayers for our troops and our president" invasion supporters are always making sure everyone knows they perform, i haven't heard them mention having so much as a moment of silence for the iraqis who are caught between the hammer and the anvil in that hellish place.

don't get me wrong: i'm not saying i'm christ-like in my concern for them. 6 days out of 7, i probably don't spend more than 5 minutes thinking about it, in all honesty. and i hate myself for that.

but nevertheless, i'm angered and confused and saddened by the fact that we seem to show far less concern for the deaths of our sisters and brothers in iraq than we show when, for example, an attractive young white woman goes missing in the united states.
posted by lord_wolf at 8:24 AM on November 6, 2004

<spin>Bah, Saddam would've killed 100,001 in that timeframe.</spin>
posted by aaronetc at 9:25 AM on November 6, 2004

I have no possible idea as to the accuracy of the figure. I somehow suspect that so reputable a journal as Lancet is not likely to put its good name on the line by concoting things. But, that said, the other night Noam Chomsky was speaking via TV monitor to panel run by Tina Brown. Chomsy used the same figure. Andrew Sullivan went crazy and shouted that Noam was a liar and it was typical of lefties to make up stuff like that and that is why America didn't support them in this election. I am not sure where Chomsky got his figures from.
posted by Postroad at 12:58 PM on November 6, 2004

I find it hard to believe the Lancet would be making this up, or that the BBC would publish it were it not without merit.

Can someone out there tell me why a different, lower number is more believable?
posted by xammerboy at 1:22 PM on November 6, 2004

Oh, come on, we all know the brown and black people that died 9.11 weren't really brown or black. They're like us, concerned with capitalism and stuff.

I saw a documentary last week where they talked to a number of middle-eastern women who have lived in the U.S. for varying lengths of time. One of them was born in the states and recalled a conversation with a friend after 9.11 where she explained that she was afraid to walk around in her own city (New York) because of the way people looked at her now. He friend turned to her and said "oh, I don't think of you as being Arab!" as if it was somehow comforting.

There are many westerners who believe they are comfortable with these "other" people, but are really only comfortable with them when they conform to their own narrow views of normality. I sincerely doubt that a plurality of Americans care how many Iraqis die. They care how many soldiers die. You know what? When I contrast eleven hundred dead soldiers with fifteen thousand dead Iraqi citizens (a conservative guess), I care a lot more about the dead in Iraq than I do about the "liberators," even if many of them didn't sign up for what it is they're doing. I also feel responsible, as a westerner, for much of the awful shit done in the name of freedom and democracy in the past sixty years. I can't imagine how much worse it would feel if I was an American.
posted by The God Complex at 3:23 PM on November 6, 2004

These weren't people, it was Satan!

"The enemy has a face. It is Satan's. He is in Fallujah, and we are going to destroy him."
Colonel Gary Brandl of the United States Marine Corps

posted by X-00 at 4:48 PM on November 6, 2004

If Satan's in Fallujah, who's in Samarra?
posted by homunculus at 5:31 PM on November 6, 2004

'==' /= '/=' unless I'm very much mistaken.
posted by stavrosthewonderchicken at 5:32 PM on November 6, 2004