October 12, 2006
Confirming or Debunking the Lancet Study with One Simple Question
The controversial and disputed Johns Hopkins study published (free reg required) in the Lancet today claims an additional 654,965 deaths as the result of the Iraq War since 2003, 610,000 of those deaths as a result of violence. It also claims they were able to verify that 92% of those 629 claimed killed in their survey had valid death certificates.
Using the research of the John Hopkins study, the Iraqi Ministry of Health should be able to therefore produce roughly 602,568 total death certificates (654,965 x 92%), and 561,200 (610,000 x 92%) of these death certificates should by attributed to violent deaths, if they do in fact collect such information nationally.
If they have far, far less death certificates on file, then the Lancet study will have invalidated itself using it's own methodology, would it not?
I'll also be very interested to find out whether or not Gilbert Burnham of John Hopkins or the editors of the Lancet made any attempt to check their figures against any available compilations of the number of death certificates issued in Iraq as a check on their research. Iraqi morgues regularly and independently released their own figures until September, when the Iraqi government took over that responsibility, which was after the data in the study was compiled by June of 2006.
Other Estimates
Not surprisingly the study figures--far beyond every other survey done by orders of magnitude--are widely discounted by most, and run contrary to every previous attempt to estimate casualties.
Iraq Body Count, a well-respected site that tracks the number of casualties in Iraq based upon media reports, lists the maximum number of casualties to date at 48,693.
The Brookings Institute reports (PDF) their estimate, based upon IBC and United Nations Cumulative data until August 31, 2006, to be a slightly higher figure of 62,000 civilian deaths due to violence. Michael E. O'Hanlon of the Brookings Institution said of the Lancet numbers:
"I do not believe the new numbers. I think they're way off," he said.
A June 25, 2006 Los Angeles Times report comes up with another set of figures:
The Times attempted to reach a comprehensive figure by obtaining statistics from the Baghdad morgue and the Health Ministry and checking those numbers against a sampling of local health departments for possible undercounts.The Health Ministry gathers numbers from hospitals in the capital and the outlying provinces. If a victim of violence dies at a hospital or arrives dead, medical officials issue a death certificate. Relatives claim the body directly from the hospital and arrange for a speedy burial in keeping with Muslim beliefs.
If the morgue receives a body — usually those deemed suspicious deaths — officials there issue the death certificate.
Health Ministry officials said that because death certificates are issued and counted separately, the two data sets are not overlapping.
The Baghdad morgue received 30,204 bodies from 2003 through mid-2006, while the Health Ministry said it had documented 18,933 deaths from "military clashes" and "terrorist attacks" from April 5, 2004, to June 1, 2006. Together, the toll reaches 49,137.
Obviously, the Johns Hopkins study figures published in today’s Lancet are far higher than any previous estimates. It will be quite interesting to see if these figures—already dismissed by every world leader and military leader commenting on it so far—can indeed be defended.
As Bryan notes at Hot Air:
The Lancet study would have us believe that 2.5% of Iraq has been killed by the war in the past three years. It would have us believe that more Iraqis have died as a result of a mid-sized insurgency than Americans died in World War II. Or the Civil War. Or Germans, who died in World War II, fighting against the combined might of the USSR, the British Empire and the United States, at a time when Germany was reduced to conscripting young boys and old men to resist those armies as they approached Berlin.This study, in other words, is nonsense on stilts.
Of course it is, but it will be most entertaining if we can debunk them using their own informaiton against them.
Update: A word on public health methodology from a medical professional at Jane Galt:
And sorry, but the defense that it's as soundly designed as can be expected for these kinds of public health surveys is a weak one. Retrospective, interview-based studies like this are poor designs. It may be the standard way of gathering data in the public health field, but that doesn't make it the best methodology, and it certainly doesn't make its statistics sound. For too long the field of public health has relied on these types of shotty shoddy numbers to influence public policy, whether it's the number of people who die from second hand smoke or the number who die from eating the wrong kinds of cooking oils.
The same blog post notes that Lancet-published studies of the past have been throughly debunked for shoddy research.
If they have far, far less death certificates on file, then the Lancet study will have invalidated itself using it's own methodology, would it not?No. Not at all. Iraq is a chaotic country with an ineffective government. There are many little fiefdoms and regions that are effectively independent, with very little coordination with the central government. It would be absolutely astounding if, with Iraq in the state it is now, there was any central government authority with a massive database of every single death certificate that had been issued by every local authority in the country. Only statistical methods, imprecise as they are, can hope to give even a vaguely accurate picture.
Iraq Body Count, a well-respected site that tracks the number of casualties in Iraq based upon media reports, lists the maximum number of casualties to date at 48,693.Can you please, just to ease the mental strain on the rest of us, acknowledge that you understand that the IBC is based on media reports, and not every single death in the country is reported by the media? You get that right? As the IBC themselves say:
Our maximum therefore refers to reported deaths - which can only be a sample of true deaths unless one assumes that every civilian death has been reported. It is likely that many if not most civilian casualties will go unreported by the media. That is the sad nature of war.Posted by: Mat at October 12, 2006 10:03 AM
Can you please, just to ease the mental strain on the rest of us, acknowledge that you understand that the IBC is based on media reports...
You mean, like in the exact section you quoted?
dee-dee-dee...
And I'm waiting on your explaination of why the Brookings and L.A.Times estiamtes are so off, and yet close to each other, and why every government leader in the free world, including the Iraqis themselves, think this report is utter bunk.
Do you really, honestly think that the earlier estimates were off by more casualties than were compiled in the the U.S Civil War, or those sustained by the German Military in World War II, and the entire world just fault missed it?
You mean, like in the exact section you quoted?What came after the ellipsis, dude? It was important, you know. The IBC is an under-estimate.
The Brookings numbers are based on the IBC.
As to what's wrong with the LA Times numbers, it's right there in the link you provided:
Many more Iraqis are believed to have been killed but not counted because of serious lapses in recording deaths in the chaotic first year after the invasion, when there was no functioning Iraqi government, and continued spotty reporting nationwide since.
...and why every government leader in the free world, including the Iraqis themselves, think this report is utter bunk.Every leader in the free world? You have a quote from the Prime Minister of Belgium?
The reason why the leaders of Iraq and the USA want to deny these numbers is that they're trying to keep control of a country on the verge of civil war, and bad news just makes this worse. And there's nothing even wrong with that. But they're hardly disinterested commentators.
Do you really, honestly think that the earlier estimates were off by more casualties than were compiled in the the U.S Civil War, or those sustained by the German Military in World War II, and the entire world just fault missed it?I believe that the earlier estimates were dramatic underestimates, mainly because the very people who compiled them claim... that they're dramatic underestimates. To be honest, I'm not entirely convinced that it's 655,000 myself, but this survey has by far the best methodology I've seen so far, so until a better survey comes along, that's what I'm going with.
And I don't know why you keep going on about this "and we just missed it" line. Which news are you watching? Every day we get news about an attack or a discovery of butchered victims. No-one's missing anything, except reliable overall statistics, and that's precisely what the John Hopkins study is supposed to correct.
Posted by: Mat at October 12, 2006 10:50 AMSo Iraq has almost the same rate of population loss as Japan had during WWII? 550 Iraqi dead per day? Give me a break. With the last study these folks did right before the 2004 election, this is beyond questionable methodology. It borders on election tampering.
G. Hamid
Could someone with more military knowledge than myself provide me with air strike figures from the period of June of 2005 to June of 2006 or tell me where I might be able to get such figures?
What I want to know specifically is how many actual targets were struck during this period?
Posted by: mike at October 12, 2006 01:41 PMI see Mat has already corrected your misuse of Iraq Body Count. However, keep in mind that there's something else Mat missed.
IBC counts *only* violent, civilian deaths recorded by two English speaking sources.
It doesn't count military actions, terrorist actions, etc., and it only counts those that are dual reported.
If you said "if you could prove to me there were more than 40,000 innocent victims of this war, I'd agree it was wrong" (I can't imagine your actually saying that...), then I could send you to IBC, and you could see that there have certainly been many more than that.
The study tried to chase down all additional deaths, from all causes, regardless of whether or not they were reported.
So, it tried to grab deaths of civilians, military, police, insurgent, terrorist, or unknown, for any cause, and then subtracted the number of deaths we would have expected, had things stayed the way they did in the 14 months pre-invasion.
As for the big numbers of deaths, 833 deaths per day is one 30,000th of the Iraqi population per day. If there were 833 deaths per year, ten people handled each body, and they all took turns, in one year, a mere 1/8th of Iraq would have had to take a turn being one of the ten people on body duty for a single body. Yes, the large number of deaths could be swallowed up, especially when it quickly stopped being "news". Do you see it in the newspaper when someone mugs a person in the bad part of town?
The study, taken over completely different populations, provided greater support for the 2004 report, coming up with similar number, with greater precision. Instead of 8-192k, centered on 98k, we have about 60-160k, centered on 112k. If you study the same phenomenon twice, and get similar results, you're virtually certain to be on the right track. Sure, your methodology could be screwed up, but screwed up so as to skew the results to such a similar estimate?
Face it; you've got strong evidence that the invasion has killed hundreds of thousands of people. Your choice is to try to learn the truth, or to insist you don't like the truth and walk away.
But until you find a *real* problem with the evidence, something more than "Bush thinks it's wrong, and the Iraqi government thinks it's wrong and (some UK officials - I don't know if Blair commented or not) say its wrong!", you're refusing to accept the strongest evidence we have.
Of course Bush would refuse to accept the evidence; he doesn't want to be known as the man whose invasion caused 650,000 deaths. Ditto for the UK, who were in it from the start, in a big way. Iraq doesn't want to admit they can't keep their citizenry safe.
But what are the statisticians who have dug into the guts of the report saying?
One of them - me, though I'm not an expert on experimental design, I can check the numbers and assumptions, and they *do* add up - says that it's solid. Rock solid. It would just about require deliberate fraud to make this report inaccurate. And if you're going to claim deliberate fraud, you ought to have evidence, or, again, you'd be refusing to accept the strongest evidence we have.
Face it; you've got strong evidence that the invasion has killed hundreds of thousands of people.
No, I do not. I have a bit of political agi-prop printed in a public health journal with a history of (a)sensationalism, and (b) publishing throughly debunked research (which I will further detail in my next post).
Again, it goes back to GIGO= Garbage in, Garbage Out.
I've read the report, and think I understand the methodolgy enough to know that it is a slightly larger sample size repeat of the same methodology used and dismissed in it's previous report of two years ago, that to date, not another single study or estimate has supported. Not one.
ThE United Nations did not support the figures in the 2004 report, and is surprised at the figures shown now. The L.A .Time estimate released in June reported 50,000 and acknowledged it was probably underreported... but not by the outrageous factor of more than ten that this John Hokpins study suggests.
You lefties keep trying to win this argument saying that the methodology is sound, without acknowledging that the researcheven admits it is far from accurate. It is a guestimate, a number arrived at by visiting separate areas that may or may not be representative, by doctors that may or may not have political motivations (remember the Fallujah doctors that swore white phosphorous was used again and again, when the bodies they showed were completely inconsistent with WP wounds?), interviewing people who also may have motivations of their own. This is a house of cards built upon a scientific-sounding framework that falls apart under an absurd casualty claim more than a half million above any prior estimate.
According to you, there are 600,000 relatively fresh civilian graves in Iraq.
Funny how no one can find them.
Posted by: Confederate Yankee at October 12, 2006 05:01 PMThe Baghdad morgue received 30,204 bodies from 2003 through mid-2006, while the Health Ministry said it had documented 18,933 deaths from "military clashes" and "terrorist attacks" from April 5, 2004, to June 1, 2006. Together, the toll reaches 49,137.
The L.A. Times numbers only take into account the deaths that occured in Baghdad, what about the rest of the country, if 50,000 deaths occured in the capital, a number of 100,000 t0 150,000 is not that far fetched for the rest of the country (Baghdad has 1/6 of the whole Iraqi population, so maybe a number of 300,000 is not that far fetched)
Posted by: Mike Huntley at October 15, 2006 01:40 PM
