Into the Valley of Death rode the six hundred thousand

The people who brought you that controversial Lancet study back in 2004 claiming 100,000 Iraqis had been killed since the invasion have revisited the numbers, with a wider sample, and come up with an even more whopping figure: more than 600,000 Iraqis killed since the commencement of major mission-accomplishment operations.

I can't make head or tail of the way the numbers were compiled, and Our Leader assures us the methodology is discredited. The Iraqi government disputes the figures as well, and as with the previous study, the margin for error does not inspire much confidence:

The figure breaks down to about 15,000 violent deaths a month, a number that is quadruple the one for July given by Iraqi government hospitals and the morgue in Baghdad and published last month in a United Nations report in Iraq. That month was the highest for Iraqi civilian deaths since the American invasion.

But it is an estimate and not a precise count, and researchers acknowledged a margin of error that ranged from 426,369 to 793,663 deaths.

Full study.

Reason's own Brian Doherty once had a great observation about how fast and loose you can play with impressively large statistics: In almost any news article, you could replace any citation of "100 million" with "100 billion" or "100 thousand" or "eleventy kajillion," and get basically the same rhetorical effect. Still, even the lowest estimate here seems pretty damn high. A few dozen more and I may have to rethink all the shameless cheerleading for the war Reason's been doing all this time.

Comments

Pro Libertate | October 11, 2006, 3:10pm | #

Didn't Reason start the war with Iraq? That's what I've always thought.

jason | October 11, 2006, 3:17pm | #

5 people dying is a tragedy, 5 million dying is a statistic... or something like that... said Stalin.

Iron Lungfish | October 11, 2006, 3:23pm | #

the margin for error does not inspire much confidence

Jesus F. Christ. The margin is essentially a bell curve; the values at the edges have the lowest probability while the values closer to the middle hae the highest probability. I realize you probably don't know anything about statistics and haven't looked at the study (PDF), but if you don't know anything about a subject you probably shouldn't make comments about how much "confidence" you have in it.

Iron Lungfish | October 11, 2006, 3:27pm | #

the margin for error does not inspire much confidence

Jesus F. Christ. The margin is essentially a bell curve; the values at the edges have the lowest probability while the values closer to the middle hae the highest probability. I realize you probably don't know anything about statistics and haven't looked at the study (PDF), but if you don't know anything about a subject you probably shouldn't make comments about how much "confidence" you have in it.

Astro | October 11, 2006, 3:36pm | #

If this paper uses the same analysis as the previous one, the error margin is the 1 sigma confidence interval, which encompasses the 67% certainty range. In my field of research, no one would publish 1 sigma results, particularly when their range is so big. The medical field is notorious for this type of sloppy analysis, though (not just when their results are politically charged).

Optimist | October 11, 2006, 3:42pm | #

Wow.

Well, let's hope those were all the BAAAAD Iraqis!

R C Dean | October 11, 2006, 3:46pm | #

I dunno, I just find it somewhat incredible that there have been an average of 500 Iraqis killed every single day since we invaded. If we are really killing Iraqis at that rate, where are the mass graves?

BG | October 11, 2006, 3:47pm | #

I think there has for a long time been a huge and growing pile of evidence (and this is just one more addition to that pile) that:

1 Shit is generally fucked up in Iraq.

2 It is not clear what is being done, or what should be done, to fix that.

I expect criticism of my post to come in the forms of:

1 "Duh, you might as well point out that rain is wet."

2 "This post is just an unproductive cheap shot at ____." (Although I am not necessarily saying anything about blame)

My reply to those sentiments is that I am just vaguely agreeing with the sense of Tim's post and not really making a strong point for or against anything. At least not yet, I'll wait for other commenters start whatever debates they would like.

JustSayin' | October 11, 2006, 3:50pm | #

If they keep hyping that number up, they will come dangerously close to saying Rummy is extremely efficient and thus winning the war. What is that, 2,500 of "us" versus 600,000 of "them?" Hard to tell a redneck that ain't winnin'.

They better tone it back down to the 100,000 range (since it is totally made up anyway).

spacekakes | October 11, 2006, 3:54pm | #

"If this paper uses the same analysis as the previous one, the error margin is the 1 sigma confidence interval, which encompasses the 67% certainty range. "

If you read the study (linked to by a previous comment), you will see that their uncertainty represents a 95% confidence level. For Gaussian statistics (bell curve) that is about 2 sigma.

Iron Lungfish | October 11, 2006, 3:55pm | #

Astro, if you're complaining that the confidence margin is too wide, then you have to note that the upper bound isn't 600,000 deaths - it's 800,000. This certainly doesn't help the "those numbers sound too high" skeptics.

BG | October 11, 2006, 4:06pm | #

JustSayin'

The goal is not and never was to kill as many Iraqis as possible irrespective of combatant/non-combatant status.

The goal was to remove the Baathists from power, prevent pro-terrorist and/or theocratic groups from taking over, and to do so with as little harm to innocent civilians and US/coalition troops as possible.

The first part was pretty much accomplished but they have been less successful with the rest of it.

Ken Shultz | October 11, 2006, 5:39pm | #

Still, even the lowest estimate here seems pretty damn high.

Really? So you think 150,000 dead people a high price to pay? ...for what we got and what they got?

Shirley there are people out there who think 150,000 a perfectly acceptable price to pay. ...but think 600,000 too much, are there not?

Are any of 'em reading this thread?

MUTT | October 11, 2006, 5:57pm | #

Well, BG, I gotta say :the media is the message.
the methods used to remove the Baath from power- and, if you remember Bush Srs remarkable lapse into honesty immediatly after Gulf War I- we are happy to do business with the Baath, without Hussien (remarkably in line with Carters "Samocismo sin Samoza" ultamatim to the long suffering Nicaraguans) the methods used in Gulf War II and the Occupation MEAN heaps of dead civilians. And Iraqis who, legitimately, resist a foriegn occupation. As did the US run embargo, that allowed Hussien to "secretly" sell oil and buy military /police equipment while insuring......how many dead old folks & kids thru disease? bad water? Hussiens well fed killers?? many tens of thousands? more??
I dont think minimizing civilian deaths was a concern for the Master Planners (AKA Serious Thinkers) but Im sure it is for Bn & Co level planners. And Plts & squaddies, too, til frustration & fear overcomes notions of doing good.
And thats the crime here. Well one of em.......

R C Dean | October 11, 2006, 6:15pm | #

So really, people are swallowing the idea that we have killed 500 Iraqis a day, every day, without letup, since the invasion?

You realize that 600,000 deaths exceeds the total killed, military and civilian, for France, or Italy, or Great Britain, in WWII? Does that really seem credible to you? That the US invasion and occupation of Iraq has resulted in more casualties than the German invasion and occupation (and Allied reconquest) of France? Than the Blitz and the land war in Asia and Europe and Africa cost the English?

Clemsonuee | October 11, 2006, 6:39pm | #

While I am not saying that I am buying the number, I do think it's high, I don't see the fact that it is larger than the number of French dead as too large, the war moved pretty fast through France both times through.

Though honestly, any of those nummbers is too much for me. I thought the war was a bad idea from the day Bush started to talk it up. I really can't see how they are going to avoid a civil war now, and that will really pile up the casualties.

Though as said before it's not the difference between 150K and 600K that will make the differnce in anyone's opinion. In all honesty it is not the number of Iraqi dead period that will form most people's judgement. It is whether the US is better off or not that will make up most people's mind. The rest will be rationalized to fit whatever opinion they reached on the first question.

I don't think we are better off, so I see any death in Iraq as a waste. The differnce between the estimates doesn't change my mind that it was a bad idea, it just means that there are more or less wasted death's.

Stevo Darkly | October 11, 2006, 6:44pm | #

I can't read the NYT article, don't understand "sigmas" no matter how many times they are explained to me, and when I click on the link to the Lancet study, it makes my browser choke. But the radio news this morning put it this way: "600,000 Iraqis killed, plus or minus 200,000."

I find this astounding on several levels.

1) Oh my God, that is a lot of dead Iraqi people.

2) Wait, isn't that a margin of error of 30%? WTF?

3) Still, it's a lot of dead people, no matter what. Jesus.

4) Assuming the numbers weren't pulled from someone's ass.

5) I don't know yet whether R C Dean's comment should make me more skeptical or more apalled.

joe | October 11, 2006, 6:54pm | #

RC,

"So really, people are swallowing the idea that we have killed 500 Iraqis a day, every day, without letup, since the invasion?"

Read the article: the range presented is for deaths from combat with Americans, but for violence in all forms, including civilians caught in crossfire, those targetted by death squads and terror groups, crime and the ongoing civil war/ethnic cleansing.

The United States lost 600,000 military along in this period of time during the Civil War, and there were no sustained campaigns against civilian populations, or serious breakdowns of law and order.

The Hutu militias killed a similar number in less than 100 days, many with machetes.

joe | October 11, 2006, 6:56pm | #

RC,

"So really, people are swallowing the idea that we have killed 500 Iraqis a day, every day, without letup, since the invasion?"

Read the article: the range presented is for deaths from combat with Americans, but for violence in all forms, including civilians caught in crossfire, those targetted by death squads and terror groups, crime and the ongoing civil war/ethnic cleansing.

The United States lost 600,000 military along in this period of time during the Civil War, and there were no sustained campaigns against civilian populations, or serious breakdowns of law and order.

The Hutu militias killed a similar number in less than 100 days, many with machetes.

Chris S. | October 11, 2006, 6:58pm | #

So really, people are swallowing the idea that we have killed 500 Iraqis a day, every day, without letup, since the invasion?

No, the study doesn't say that "we" have killed 500 Iraqis a day. Most of these alleged casualties were apparently the result of Iraqi infighting/civil war (or whatever you want to call it) and poor living conditions subsequent to the invasion.

You realize that 600,000 deaths exceeds the total killed, military and civilian, for France, or Italy, or Great Britain, in WWII? Does that really seem credible to you? That the US invasion and occupation of Iraq has resulted in more casualties than the German invasion and occupation (and Allied reconquest) of France? Than the Blitz and the land war in Asia and Europe and Africa cost the English?

No, that's not a fair comparison. You're talking about direct casualties and the study includes deaths indirectly caused by the war. Moreover, you're cherry picking. You've specifically excluded the holocaust and Russian casualties, which completely changes your WWII analogy.

I have no idea whether 600K is accurate. The study is plainly politically motivated, and it appears to be based on some questionable assumptions. But let's try to be honest in our criticism.

spur | October 11, 2006, 7:02pm | #

"So really, people are swallowing the idea that we have killed 500 Iraqis a day, every day, without letup, since the invasion?"

The study doesn't say The US killed them, just that they got killed, that includes personal murder (domestic violence, etc.) The study also doesn't include the invasion, just occupation onwards as I understand it. Yes, Iraq was less violent under Saddam even if the number was 80,000...

Chris S. | October 11, 2006, 7:02pm | #

So really, people are swallowing the idea that we have killed 500 Iraqis a day, every day, without letup, since the invasion?

No, the study doesn't say that "we" have killed 500 Iraqis a day. Most of these alleged casualties were apparently the result of Iraqi infighting/civil war (or whatever you want to call it) and poor living conditions subsequent to the invasion.

You realize that 600,000 deaths exceeds the total killed, military and civilian, for France, or Italy, or Great Britain, in WWII? Does that really seem credible to you? That the US invasion and occupation of Iraq has resulted in more casualties than the German invasion and occupation (and Allied reconquest) of France? Than the Blitz and the land war in Asia and Europe and Africa cost the English?

No, that's not a fair comparison. You're talking about direct casualties and the study includes deaths indirectly caused by the war. Moreover, you're cherry picking. You've specifically excluded the holocaust and Russian casualties, which completely changes your WWII analogy.

I have no idea whether 600K is accurate. The study is plainly politically motivated, and it appears to be based on some questionable assumptions. But let's try to be honest in our criticism.

BG | October 11, 2006, 7:43pm | #

MUTT

Well, I suppose I don't have first hand knowledge of the motivations or intentions of the president or any US leaders. I don't know exactly how much of a priority it was to avoid/minimize harm to civilians. From what I understand however, the number of civilian casualties due to US munitions (as well as probably the rate of civilian casualties measured per bomb or per unit of time or other reasonable measures) was less during the invasion than either the first gulf war or most other comparable conflicts.

If I have my facts straight, the vast majority of Iraqi civilian casualties occurred after the baathist regime fell and were killed/injured by either death squads or militant forces outside of (and usually opposed to) the occupying coalition. Those included many accidents and many incidents of deliberate targeting.

The point I was making to JustSayin' was that the goal was not to try to maximize Iraqi casualties (as his comment implied). If that actually is what our president or other leaders are trying to do; then it is clear that they are not only repugnant but highly incompetent as well, given the nature of the arsenal at their disposal.

I know that Bush senior had no interest in going on to Baghdad and forcibly removing Saddam or the Baath party. His stated reason for not doing so was pretty much that it would result in the kind of situation we have now. It now seems that may have been a sensible position on Iraq (although there were good reasons to object to the sanctions initiated during his term).

MUTT | October 11, 2006, 8:13pm | #

Correct, BG- there was no notion to MAXIMISE civ casualties. There just was no thought to minimize them: Iraqi life simply isnt important.
I was just listening to one of the authors of the "study" on the radio. - A word that should always be placed in quotes.
Im not a statistician, its an arcane art, and I think, easily bent ("Lies, damn lies, & statistics") According to this person, whos name escapes: as above, a bell curve. Greatest chance of accuracy towards the middle of the bell; method-door to door canvassing, in "many" different locales/stratas, asking to see death certificates; and going to cemetaries to inquire about thier records.
And its not like the US hasnt in the past desired to maximise casualties, so there you go. Why else would it be the stock response to "how many civ casualties" is "we dont keep those records" by spokesmouths.
Ah- you say: prove it. OK. No longer secret studies on how long it took for first responders to start putting out fires & searching for survivors after air raids in N VietNam. 2 hours. So, thats when the second wave came in. Historical, public record. Look it up.
What Im saying here is Ive never noted any US reluctance(in my lifetime) to inflict massive civilian casualties. And as I point out with Bush Sr's desire to go back to business with a Saddam-less Baath, no reluctance to get in bed with mass murderers.

Retief | October 11, 2006, 9:02pm | #

600,000 isn't that many. Look. Iraq's got a population of 26 million. The pre-invasion deathrate that the study found was ~5.5/1000 That's right in the neighborhood of US government, UN, and everybody else's estimates of mortality in Iraq under Saddam. The CIA world factbook estimates the deathrate in Iraq at 5.37/1000. 5.5/1000 times 26 million is ~143,000. So over the last 3 and a half years since March 2003, if the death rate hadn't changed one bit (as the CIA suggests) we'd expect 500,500 Iraqis to have died from all causes. So baseline is 400 Iraqis dying a day with no war at all.

The study in question found that, shockingly, death rates were higher for the period including the invasion and the current troubles than they were for the period before that. (With a war going on is this a surprise?) They did it by choosing as sample of housholds and counting the people that had died in those housholds during the two periods. The result for the prewar period matches closly with other estimates as I mentioned. They found that the post-invasion deathrate was 13.3/1000. In the households they studied the deathrate more than doubled, expand that out to the population of 26 million and you have an "extra" 600,000 deaths. It's not that many when you consider how many people there are in Iraq.

Now maybe people sample didn't remember the folks in their houshold who died pre-march-2003 as well as the more recent ones or maybe the 13.3 is an overestimate of the post invasion deathrate for other reasons. Maybe the current violence, lack of medical care, and extra stress, plus the invsion deaths only doubled the deathrate versus previously. That would obviously give us and additional 500,500 deaths in Iraq versus the baseline. (And yes that's just additional deaths without regard to innocence or guilt) Or maybe it is way overestimating the increase in mortality resulting from the current violence plus the invasion. Maybe that only increased the deathrate in Iraq by 50%. Does that sound reasonable? That would mean 250,250 dead Iraqis from the increase. Still a lot. But none of these nubmers are that big in the context of 26 million people in Iraq and three and a half years since the invasion.

On the other hand to believe that 30,000 is the right number you have to convince yourself that the invasion and continuing violence in Iraq changed the death rate there by a meager 5% from 5.5/1000 to 5.8/1000. The entire effect on mortality in Iraq of the invasion and the violence over the three and half years since then is .3/1000 or an increase in the number of deaths of 3 people per 10,000 people? Now what number sounds outlandish?

Now try forgetting all that and see if you can imagine what effect on the deathrate in Iraq you might guess the invasion and the violence might have had before you knew these numbers. Maybe your answer isn't "doubled" but I'll bet it's not "barely moved" either. Now multiply by 3.5 years and 26 million people and you'll find that you too believe that more than 100,000 more people have died in post-invasion Iraq than would have had the baseline deathrate continued.

spur | October 11, 2006, 9:27pm | #

Retief,

Not a big number in the context of 26 million? If those numbers were switched to the US, you'd see an increase in the deathrate of (somebody do the math for me) several million people.

And then we'd read European or Iraqi bloggers going: 'well in the context of 300 million people, a couple million more violent deaths isn't that bad.' And we'd be telling them to piss off.

joe | October 11, 2006, 9:42pm | #

Regief,

Read the intro to the study: the 600,000 represent the number of deaths above and beyond the 5.5/1000 mortality rate prior to the invasion. This represents at least a doubling of mortality in Iraq since the invasion.

Let me put that another way: think of your chances of dying if you lived in Saddam Hussein's Iraq. Now double that, with all of it coming from an increase in your chance of dying violently. That is what has happened in that country since Dick Cheney decided to liberate them.

How many of those hundreds of thousands of additional homicide victims had purple fingers, I couldn't tell you.

BG | October 11, 2006, 10:30pm | #

MUTT

It is true that there are many historical instances in which the US government aided or aligned with despotic regimes with horrible human rights records. It is also true that there are cases in which US leaders deliberately acted to maximize harm to innocent people or acted without regaurd to such harm. In some cases, I would not be opposed to prosecuting individuals responsible for such acts who are still alive.

It seems to me though that the tendency for US military operations to cause innocent casualties has been decreasing in recent years. This is partially because they have more accurate targeting systems and partially because people didn't care as much back then.

However the examples you cite do serve to point out a problem: there seems to be little in the way of effective checks or limitations on how the people in charge of the US government and military apparatus use that power abroad. There are treaties and international law, but those seem to have limited effectiveness. And we could imagine a malevolent future administration witdrawing from those treaties and committing untold atrocities. To address this, I propose a constitutional amendment holding the US government to certain standards of human rights in its foreign policy. I don't have the details worked out but the limits it sets should take into consideration both the humanitarian impact of the government's actions as well as the strategic importantce of certain actions. I would imagine it would also contain a formal mechanism whereby US courts (which tend to have more pull with the US government than international courts) can rule on whether its provisions have been violated and what to do about such violations.

M | October 11, 2006, 11:57pm | #

other commenters start whatever debates they would like.

I find it suspect that Reason staffers never double-post in the Comments section.

Retief | October 12, 2006, 1:30am | #

spur, My point isn't that the number of 600,000 additional dead in Iraq is not a big deal. It is. The point is that the number is perfectly plausible given the population. I don't think we disagree.

Joe, I think I just said that. At length.

Stevo Darkly | October 12, 2006, 2:35am | #

I find it suspect that Reason staffers never double-post in the Comments section.

That might be because when they do, they have the ability to delete the extra post.

Sometimes I've seen them double-post topics (or whatever I should call the offical posts that actually start a thread) in the blog section, to be corrected shortly after. Recently I thought I even saw a topic triple-posted, but don't quote me on that.

John P | October 12, 2006, 8:14am | #

MUTT: Im not a statistician, its an arcane art, and I think, easily bent ("Lies, damn lies, & statistics") .

Idiotic comment. Statistics is rigorous mathematics. It is not art. It is not "easily bent." It may be difficult to understand for some. This is irrelevant to to the mathematical correctness of the analysis in the paper by Burnham et al. Your comment reminds me of the stupid political dismissal of sampling theory by Republicans when the census wanted to use it. Politicians rely on sampling theory every time they refer to a poll!

There's even an online course if you want to start learning.

R C Dean | October 12, 2006, 10:04am | #

Read the article: the range presented is for deaths from combat with Americans, but for violence in all forms

I realize that, but aren't all the excess deaths our fault for invading Iraq in the first place?

No, that's not a fair comparison. You're talking about direct casualties and the study includes deaths indirectly caused by the war.

So the study includes "speculative" deaths but still assigns blame to the Americans? Interesting.

Moreover, you're cherry picking. You've specifically excluded the holocaust and Russian casualties, which completely changes your WWII analogy.

Actually, no. The casualty figures for France and Italy include the relatively minimal holocaust casualties they suffered.

I realize the Russians lost millions. I wasn't comparing the Iraqi war casualties to WWII in toto. Can't you read?

I was merely pointing out that it seems incredible to me that more Iraqis have died in, what, three years as a result of the war than Italians, French, or British died in five years as a result of WWII.

Inherently incredible and politically explosive results released as an October surprise deserve a little skepticism, no? Yet most people commenting here are swallowing them whole.

R C Dean | October 12, 2006, 10:14am | #

As a further observation, I might note that the Iraqi society is sufficiently advanced that nearly everyone who dies there gets a death certificate. And you know the totalitarian quasi-Nazi Hussein regime kept records on who died.

Why do statistical sampling, when you can do what would probably be a pretty damn accurate actual count of the deaths pre- and post- invasion? Why ask people what their recollections are of who died and when, when you can go look it up?

Factory | October 13, 2006, 12:54am | #

One should also note that Australia's death rate is 7.51 and the US's is 8.26. If Iraq's death rate is 5.5 pre-invasion, erm it doesn't sound all that bad, and 13.something erm, is less than double that of the US's.