« Hacktacular | Main | A Promissory Note »

Lancet Update

Daniel Davies makes a good point regarding the extraordinarily large margin of error surrounded the Lancet study's estimate that the war has caused 100,000 deaths (give or take 92,000) above what would have been expected had the invasion not taken place:

The critique here, from Slate, is that the 95% confidence interval for the estimate of excess deaths (8,000 to 200,000) is so wide that it’s meaningless. It’s wrong. Although there are a lot of numbers between 8,000 and 200,000, one of the ones that isn’t is a little number called zero. That’s quite startling. One might have hoped that there was at least some chance that the Iraq war might have had a positive effect on death rates in Iraq. But the confidence interval from this piece of work suggests that there would be only a 2.5% chance of getting this sort of result from the sample if the true effect of the invasion had been favourable. A curious basis for a humanitarian intervention; "we must invade, because Saddam is killing thousands of his citizens every year, and we will kill only 8,000 more."
I, for one, don't think the humanitarian argument for war really needs to be taken seriously, since, in my opinion, it's rather obviously offered in bad faith. Daniel, however, has long been interested in the point (I assume that in the Blair-led UK it's a more salient topic) and the Lancet study really is devastating on this point. It makes it very hard to say how many civilians have died as a result of the war, and for various reasons I think it's likely much closer to 8,000 than to 100,000, but a well-conceived charity venture doesn't wind up killing the people it was supposed to help.

November 1, 2004 | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Lancet Update:

» The Lancet report revisited from Peace, order and good government, eh?
Last week there was a flurry of media stories, and blog posts, in response to news of a study to be published by the British medical journal The Lancet that had used a combination of interviews on the ground in... [Read More]

Tracked on Nov 1, 2004 9:31:27 PM


Feh. The war isn't over. And nobody in their right mind expects a war to produce positive effects while you're still fighting it. The real test, as anybody sensible understands, is how Iraq is doing five or ten years from today.

Posted by: Brett Bellmore | Nov 1, 2004 6:15:54 PM

I think I agree with Brett. The relevant figure doesn't seem to be how many people died since the war started but how many will die over the next five-plus years.

Posted by: Luka Yovetich | Nov 1, 2004 6:47:22 PM

Bush has objectively saved over 100,000 Iraqi children since he liberated Iraq: At Least 650 Billion Civilians Dead In Iraq.

Posted by: Modern Crusader | Nov 1, 2004 6:49:28 PM

Actually, I disagree with Brett. The relevant figure is not the number of lives to be saved over the next 5 or 10 years. Rather, it is the number of lives Saddam took over the past 25 years. And, of course, THAT number is conservatively estimated at about 300,000 Iraqis (and that, of course, doesn't even count the number of Kuwaitis and Iranians, nor the number of Iraqi soldiers killed, in the two wars Saddam started).

Left-wingers have got this asinine idea that if Saddam was slowing down his killing machine over the few years just prior to the war, we ought to forget about the hundreds of thousands of people he killed in the 20 years before that. Which makes no sense. Saddam killed hundreds and hundreds of thousands of people. Dwarfing by an order of magnitude even the unbelievable numbers published in the Lancet.

Oh, and, of course, reason #862 not to believe the study published in the Lancet - it was only published on condition that it be released before the election. The study is obviously bogus.

Posted by: Al | Nov 1, 2004 7:07:34 PM

Nah, those are sunk costs, Al. If Saddam had definately stopped killing people, and the only way to bring him to justice was to kill a bunch of innocent people, it wouldn't be worth it. It was worth it only because,

1. He hadn't stopped killing people.


2. Most of the people we're killing aren't innocent.

Posted by: Brett Bellmore | Nov 1, 2004 7:30:52 PM

Bret Bellmore:

The war may not be over (probably never will be under Bush) but major combat operations are over, so what you see is what you get. Remember, the Republicans have been doing POLLING in Iraq to prove that things are better there. This is certainly more accurate than that.

After being responsible for a million deaths in Vietnam, we pulled out, leaving the place worse than when we got there. The same will be said about Iraq. Pointless. Every death a crime, and our national interest suffers. Spin on Brett.

Posted by: epistemology | Nov 1, 2004 7:31:28 PM

Poor Al, just losing it at every turn.

Brett's argument is at least an argument; what Al has typed here is piffle.

No, we don't have to forget about the genocide that Saddam was willing to carry out when he was our ally; we don't have to forget about the genocide he committed after we encourage an uprising against him and then walked away. But neither do we have to act as though genocide were still going on in iraq in recent years when it wasn't. The idea that we should expend massive amounts of blood and treasure just in order to put saddam on trial for genocide is not one that the american public would have endorsed in advance of the adventure.

As for the publication, sheesh, Al, two simple points: a.) the reason that one of the co-authors wanted the study published before the election was to force both candidates to take a stand on the costs of urban warfare. this was, of course, a completely unrealistic expectation on his part, but still, it wasn't directed at bush (do you really think that bush supporters actually care enough about the iraqi people that they would vote against bush at this late date simply because iraqi civilians have been killed since our adventure began?); and b.), if you would stir yourself to look at the Lancet, you would discover that there is a standard procedure they employ for fast-tracking papers for publication. Here, i'll do the heavy lifting for you, since your synapses really aren't working so well today:

For papers, which will usually be primary research (Articles, Mechanisms of Disease, and Research Letters), judged to warrant fast dissemination, The Lancet will publish a peer-reviewed manuscript within 4 weeks of receipt. The essential first step is an e-mail to fasttrack@lancet.com or a telephone call to one of the editorial offices (London +44 [0] 20 7424 4943 or New York +1 212 633 3810).


Peer review

Every Article, Mechanisms of Disease, Research Letter, Case Report, Hypothesis, and review papers of all types that are published have been peer reviewed. Occasional contributions (eg, an Essay) are accepted without peer review. Your report will first be read by one or more of the journal's staff of physicians and scientists. Our acceptance rate overall is less than 10% and it is an important feature of our selection process that many papers are turned away on the basis of in-house assessment alone. That decision will be communicated quickly. More positive in-house views are followed by peer review by at least two reviewers for all research papers and for most other types of paper. You will receive notification of which editor is handling the peer review of your paper. If reports are encouraging, and the editorial consensus is also favourable, then statistical advice is sought where appropriate.


As i said with respect to this matter yesterday, i have no dog in this fight. I am not a statistician; i am not a public health issues scholar. The authors of the paper understand the speculative nature of their endeavor, and frankly, they could well be wrong and you couldn't prove it by me.

But the simplistic notion that this was rushed into print and is therefore "obviously bogus" is, of course, "obviously bogus."

Care to cite any of your other 861 reasons for considering the paper "obviously bogus?" maybe one of them is right, simply on the shakespeare and monkeys theory.

Posted by: howard | Nov 1, 2004 7:33:50 PM


The problem, of course, is that when Saddam was slaughtering those people, we did nothing to stop him? Remember? We attacked Saddam in Gulf War I, not after the world was shown proof of Saddam's using poison gas to slaughter innocents, but after he took OILWELLS from the dictators in Kuwait. Remember? So now Bush II, after Islamic fundamentalists have attacked us, is turning the country over to an Iranian born Ayatollah. This is good how?

Of course the US installed Saddam's Baathist party in Iraq to begin with. And why are we arguing about this anyway; it's the Wahhabi, not Baathist, version of Islam that is so pernicious.

Posted by: epistemology | Nov 1, 2004 7:42:59 PM

So, we can't clean up the messes we've made ourselves? THAT makes a lot of sense... And in any case, the US isn't a person, you're talking about actions taken by different administrations. Why shouldn't one administration take out a bad guy another played some part in building up?

Posted by: Brett Bellmore | Nov 1, 2004 7:55:47 PM

Brett, no reason at all on the merits. This came up because Al wants to count all deaths caused by Saddam at any time on the Saddam average annual deaths caused scoresheet.

Dismissing Al and attempting to talk seriously, the issue isn't that we shouldn't clean up a mess that we contributed. The question is whether the cost of cleaning up a mess that we contributed to outweighs the benefits of the cleaning.

As Wolfowitz noted long ago, and as the bush administration demonstrated through its actions, simply getting saddam because of past genocide didn't pass the cost-benefit analysis question, which is why the wmd question and the connection to al qaeda question became the admin's talking points for this war.

Posted by: howard | Nov 1, 2004 7:58:55 PM

And where the hell is Modern Crusader to put Al and Brett in their places by arguing that Saddam should have been given a medal for all the Sand Nazis (as his blog calls them, apparently too much of a pandering pussy to call them by the name he uses in private: "sand niggers") that he has killed.

Right Modern Crusader?

(Sorry to all for feeding the trolls.)

Posted by: epistemology | Nov 1, 2004 7:59:47 PM

I really don't understand the "you snooze, you lose" response to genocide. I'm really quite curious about it: how long does it take after a genocide is "concluded" until the dictator responsible for it becomes immune from attacking him for it?

I also disagree with the "sunk costs" theory. Analogize to crime. You know, once a murderer kills somebody, the corpse is likewise a "sunk cost" -- why do we expend the resources to go after the murderer? It's not solely because we are worried about him committing future murders...

Now, I accept that genocide in and of itself is not always sufficient reason for war. There are other factors to consider, including how many additional people will die in the war. Nonetheless, I consider it sufficient in this case, regardless of what Wolfowitz said.

This came up because Al wants to count all deaths caused by Saddam at any time on the Saddam average annual deaths caused scoresheet.

And, frankly, I'm not interested in the "average deaths" scoresheet. 300,000 people is 300,000 people, regardless of whether it was all at once 20 years ago or exactly 15,000 per year for 20 years. To me, that is a distinction without a difference.

Posted by: Al | Nov 1, 2004 8:12:26 PM

Well, Al, i'm sure you were busy criticizing the bush administration for not invading iraq on January 21, 2001. As for sufficiency, well, it's perfectly clear that we would never have entered this war were the only justification saddam's past genocide, so your willingness to enter this war on that behalf is fairly meaningless. (By the way, Al, why are you wasting your time on this matter instead of addressing current genocide? why aren't you pissed at bush for failing to address it?)

Meanwhile, nice avoidance of the point (although, in your own way, you've conceded the point): no, we don't act to curb every case of genocide. In fact, we rarely act to curb genocide, so the justification for acting needs to be more than that genocide has occurred in the past.

Posted by: howard | Nov 1, 2004 8:19:33 PM

Right. Pass one mugging on the street, and you will forever require extra justification to intervene if you see another. It's not just enough to say that you're finally cleaning up your act.

Posted by: Brett Bellmore | Nov 1, 2004 8:56:07 PM

50,000? 100,000. How many is too many? Our airforce is bombing Iraqi cities on a daily basis. Is there anywhere in a densely populated city we can bomb and not kill civilians? We are trying to win a political fight with military force. It never works.

Posted by: bakho | Nov 1, 2004 8:58:47 PM

brett, i can't tell what you're trying to prove with your analogy, other than that arguing by analogy isn't always the most useful way to go.

Every choice we make carries costs and benefits; is it your opinion that when it comes to addressing past genocide, we should act as though costs don't exist and virtue is its own reward?

Posted by: howard | Nov 1, 2004 8:59:57 PM

"And in any case, the US isn't a person, you're talking about actions taken by different administrations."

Technically different administrations but many of the same people: Rumsfeld, Cheney, Perle, Wolfowitz. So Rumsfeld giveth WMD's and Rumsfeld taketh them away? It would be nice if it were that simple, but the problem is these guys had bad motivations BOTH times. As has been mentioned on this thread, these same guys who turned a blind eye to genocide the first time around in exchange for a ruthless ally are now wailing about terrible regimes when it means they can control basing, oil, and government contracts. If a few hundred thousand Iraqis pay the ultimate price, who are WE to stem the tide of PNAC -- er...freedom?

That doesn't bother you guys even a little bit?

Posted by: Windhorse | Nov 1, 2004 9:09:05 PM

Note that the Lancet study is "excess mortality". That is going to be larger than just the number of bodies in mass graves. I would strongly suspect that the excess mortality from Saddam years of misrule is a lot more than 300,000. (If a child dies of, say, whooping cough because he is malnourished and has poor medical care because of war or social unrest, that is excess mortality, even though there may be no violence directly involved.) As an example, about 680,000 people were shot in the Great Purge in the USSR (if you believe the NKVD records), while the excess mortality from USSR census records during the 1930's is estimated to be 5 to 10 million people, or 10 to 20 times larger. From that sad experience, an excess mortality from Saddam's reign of 3 to 6 million would not be too surprising.

I would be very careful conflating excess mortality and casualties. It is not the same as casualties, which is what people typically care most about.

One simple way to get a gross estimate of casualties is through historical experience. In Vietnam, for example, there were about 20 civilian and 20 enemy deaths for each American soldier who died. This seems fairly typical for counterinsurgency type warfare, and would predict about 20,000 civilian and 20,000 insurgent casualties in Iraq to date. The web site
http://iraqbodycount.net/ (which lists identifiable civilian casualties) has a current civilian casualty estimate of 14219 to 16352, which is precise but incomplete, and may lump in some insurgents, but fits in well with that 40,000 estimate.

So all of these numbers hang together reasonably well (from a physicist's perspective of first trying to get a rough factor of 2 estimate). It seems clear than civilian casualties are in the one to several tens of thousands range, insurgent losses are presumably at the same level, and a roughly equivalent or larger amount of excess non-military mortality (largely from poor public health) is likely.

Posted by: Marshall | Nov 1, 2004 9:17:36 PM

Deaths from malnutrition and lack of medication were estimated at 4000-5000 per month due to diversion of the goods from the oil-for-food program. Naively, that would be about 80,000-100,000 innocent dead were it not for the war. That assumes the sanctions and the corruption continued to this day. However, war was not necessarily the only way to eliminate that corruption. Or, if war were necessary, it does not justify the incompetent prosecution of the war which has resulted in the insurgency. Had the war been fought competently, there would be less close combat in urban areas, and less innocent dead.

Also, saving a life does not justify taking lives. If a paramedic saves a life, he does not earn a license to kill. Saving Iraqis does not earn us the right to kill Iraqis.

Posted by: Njorl | Nov 1, 2004 9:20:27 PM

Once again, with feeling:

I and a lot of people don't view this as an argument that there should have been no war. (In case that's what you're wondering, Al) Rather, it's one more argument that this war has been handled incompetently from day one. It all comes back to troop levels and security concerns. If Armoured division personnel have to be used for police duties because of a lack of lighter trrops, if weapons depots cannot be secured due to a lack of personnel, if you refuse to have any foreign forces in place except under US command--these are all factors that drive that civilian casualty figure up.

Saddam was a murdering bastard. I'm glad he's done. But that has very little to do with the argument at hand, I think.

Posted by: Wrye | Nov 1, 2004 9:31:04 PM

We must punish Saddam for killing so many innocent Iraqis! Even if it means killing even more Iraqis in the process! Anyone who disagrees just doesn't care about the lives of Iraqis!

Posted by: JP | Nov 1, 2004 10:14:40 PM

Just in case anyone was sufficiently imprudent as to believe him, most of the claims in howard's first post are historically incorrect mendacious propaganda.

Don't believe crap from people like him - go off and do your own research.

Posted by: am | Nov 1, 2004 11:46:23 PM

Shorter am: Liar, liar, pants on fire!

I mean, everyone knows the Lancet study just has to be a lie. After all, if it were accurate, it would undermine the case for war!

Posted by: JP | Nov 1, 2004 11:52:30 PM

am writes:

Just in case anyone was sufficiently imprudent as to believe him, most of the claims in howard's first post are historically incorrect mendacious propaganda.

I suspect this historically incorrect mendacious propaganda issues from the same type of people who'll tell you Iraq had no WMD. Thank god we can ignore them, with their tiresome "facts" grounded in "reality."

Posted by: A Tiny | Nov 1, 2004 11:59:31 PM

Just in case anyone was sufficiently imprudent as to believe him, the claim in am's last post is incorrect mendacious propaganda.

Don't believe crap from people like him - go off and do your own research.

Posted by: Oooh, Upping the Ante | Nov 2, 2004 12:03:35 AM

The comments to this entry are closed.