Dsquared draws our attention to this study in the New England Journal of Medicine, which suggests that 13% of US Army soldiers in Iraq believe themselves to have killed a civilian.
Very back-of-the-envelope-ishly, dsquared works out that this involves 60,000 deaths *directly* attributable to coalition troops (before you take into account the rising mortality rate from disease and lack of medical care) – higher than in the Lancet study. While this extrapolation is far too rough to actually make that claim, it certainly provides significant support for the order of magnitude of excess deaths claimed in the Lancet study.
Inevitably, the more ignorant pro-war-ites will critique this study as only they know how ("but they only interviewed 894 soldiers! That means there might only be 116 soldiers who’ve killed any civilians at all!", etc), while the less ignorant ones will quietly ignore it.
This is unfortunate: everyone on whatever side needs to accept that the Iraq war (irrespective of how noble its aims and irrespective of Saddam Hussein’s undeniably great evil) has been a massive fuckup and is something that we should never, ever do anything like again.
Update: Peter C points out that if you support the war for political/strategic reasons and don’t care about what happens to the Iraqis, then you can still believe it was a good idea. Fair play.
It was a good spot from Chris via Lenin and the SW … the obvious objection to the calculation is that not all troop-days are created equal and you would expect that March 2003 accounted for all of the incidents. On the other hand, the actual data, as reported in the Lancet study, don’t bear this out; there are a lump of violent deaths in March, but there are actually quite spread out over the postinvasion period, supporting the US Army claim that they were very careful about civilian casualties during the invasion (though they’ve been getting steadily more careless as time moved on)
Hahahahaha! John’s spam filter catches the words S*****ist W***er!
Inevitably, the more ignorant pro-war-ites will critique this study as only they know how ("but they only interviewed 894 soldiers! That means there might only be 116 soldiers who’ve killed any civilians at all!", etc), while the less ignorant ones will quietly ignore it.
There’s a difference between quietly ignoring something because you feel it’s somehow destroyed your view and not having anything particular to say about it. It’s not exactly a huge shock to have a death toll in a war, and I don’t have a fixed number of Iraqis beyond which the Iraq War would be wrong.
You also seem not to have realised that those who supported the Iraq War for realist or "assertive nationalist" reasons may find these figures confirm their view most of all, because it helps build a picture of waging war for humanitarian reasons being counter-productive.
It’s not exactly a huge shock to have a death toll in a war
This many civilians, in a war of this type, with such a preponderance by violence, is actually really very unusual.
Well, unlike the number crunchers, I’m really rather more interested in the exact question asked and how it was answered (but I’m too lazy to examine closely the typical bone-headed study we get from academics today).
Whether one has killed a civilian is doubtless a difficult judgement to make in the kind of situation we’re talking about. How were the respondents asked to view this ambiguity? When, for instance, is a boy a combattant, when a civilian? What does ‘believe themselves to have killed’ mean in this context? Dsquared used the term ‘personally responsible’, but even then given the nature of warfare it’s a massive judgement to make- even if were were talking only about how many soldiers had actually caused the death of a person. One bullet from one gun is rarely the case, is it?
The survey (linked above) asked soldiers whether they’d experienced the following things (among many others):
"being responsible for the death of an enemy combatant"
"being responsible for the death of a noncombatant"
414 soldiers said yes to #1, 116 said yes to #2. The figure for Marines to #2 was substantially higher. The questions were asked in that order.
Heh – I realise the ‘UPDATE’ summary was meant to be a little crude, so I’ll not be so humourless as to take issue with what you have said. But the important point is what you haven’t. Let’s take two people who are pro-war, but for very different sorts of reasons.
There is a realist type – let’s call him Henry K – who thinks war is sometimes necessary and preferable if it enhances national security and advances the national interest, but is very sceptical about the idea of invading countries for human rights reasons. And there is an ideological, moralist type – let’s call him Tony B – who rather gets a thrill out of the idea that he can make the world a better place by invading nasty countries and saving its people.
The point is not just that a high death toll will do more harm to Tony B’s case, while Henry K’s will be untarnished. It’s also that Henry K’s case is actually advanced by a high death toll, because he can then say "Ahh, well, look what happens when we try to be nice! Better to wage war only when the needs of the nation really demand it."
Yes the stance "let’s only go to war when we really have to" is indeed strengthened by a high death toll. But many (even most) people would agree with that stance. Your Henry K however (presumably) also holds another opnion, namely "we really had to go to war with Iraq", and this position is not strengthened by a high death toll. Admittedly it’s not weakened either, but it makes a full justification for that second opnion even more important. If a convincing justification is not forthcoming (which we can argue about, in my view it isn’t) then it makes Henry K’s over all position less tenable.
That’s not a great poll question, though, is it? ‘Being responsible’ covers a multitude of sins. Seeing someone get killed by a bomb going off, a soldier might hold themselves responsible, having failed to save that person. Doesn’t mean they pulled the trigger. Equally, a group of soldiers could witness a single event, all hold themselves responsible, and hence double counting. So I’d quibble on the 60k as an ‘exact’ figure, but maybe the order of magnitude is about right. I would also quibble with your wording ‘*directly* attributable to coalition troops’ for the same reason.
If all anti-Coalition troops wore uniforms, the figures would be different: fewer civilians would have been killed in the first place, and fewer soldiers would erroneously believe they’d killed civilians. There is, therefore, an argument that we should never go to war against people who use civilians as human shields, but then we’re faced with the twin problems that (a) such people are exactly the sort of people we should fight and (b) if we let it be known that we won’t wage war against people who use such tactics, then we encourage more people to adopt them.
Andrew’s right about double-counting. What if four soldiers open fire at one man and kill him? I doubt that sort of thing’s unusual, and that’s one death being counted four times already.
> That’s not a great poll question, though, is it?
Depends what you want the poll to show. I suspect that it’s a very good poll question, and that a lot of thought went into it.
(As I’ve said before, I don’t believe there’s a magic number of deaths above which the war turns from right to wrong.)
The survey was one on PTSD among soldiers, not on Iraqi casualties, so it may be pushing it to suggest the researchers were trying to get a high civilian casualty figure. The casualty data is an extrapolation by dsquared, rather than anything published in the research.
I’d’ve said a better conclusion to take is "anyone sensible against whom we’re fighting a war will use civilians as human shields, therefore we should never go to war unless we’re confident that *given that fact*, the outcome will be still be better than no war".
How could you phrase a better (i.e fairer) question?
No phrasing could avoid the double counting issue: if 4 soldiers shoot 1 guy then any question will either yield four "yes" results or four "no" results, both of which are misleading.
" ‘Being responsible’ covers a multitude of sins."
I take this point to some extent, but I doubt many soldiers (who know what directly "being responsible" for someone’s death means) would answer "yes" to this on account of people they failed to save. Anyway if the soldier says they’re responsible for someone’s death, then who are we to argue?
In any case every "yes" to Q2 still implies the death of a non-combatant which deserves to be counted.
The Lancet Study actually had a massive impact on pro-war loons. Before it appeared it was generally implied that the "Iraq Body Count" survey was the work of Saddam Hussein himself, making up civilan casualties thousands of times higher than they really were in an attempt to discredit the Coalition and destroy the Free World from within. After the Lancet appeared it was generally implied that the "Iraq Body Count" survey was the finest piece of statistical research ever carried out, utterly impeachable in its accuracy and intentions.
If this other report gains wider publicity then we’ll probably all agree it was pretty bad sooner or later.
who are we to argue?
Armchair generals, of course.
In any case every "yes" to Q2 still implies the death of a non-combatant which deserves to be counted.
Not so. To take a ludicrous example: if a squad of 60,000 soldiers all fired one bullet into a house, then they all went in and found one dead Iraqi civilian, they could all legitimately answer "yes" to Q2. Result: 1 dead Iraqi. Our survey says: 60,000 dead Iraqi’s.
Please note that the last paragraph contained the word ‘ludicrous’, but double counting is certainly an issue here. Although as I said above, I wouldn’t dispute the order of magnitude of the estimate.
I agree with Squander Two here – there is no tipping point number of deaths which makes the war right or wrong. This is the problem with the humanitarian argument for war.
Our survey says: 60,000 dead Iraqi’s.
Look no. I really am not going to have the NEJM subjected to the kind of ignorant politicised bollocks that the Lancet brought on itself. The paper was about PTSD in American soldiers. The extrapolation of 60,000 "responsibilities for noncombatant death" was my back-of-the-envelope calculation to see whether the order of magnitude was the same. Please, people let’s be careful about this and not let it take on a life of its own.
there is no tipping point number of deaths which makes the war right or wrong
Of course there is. If we killed every single person in Iraq, then the war would be wrong. If we killed nobody, it would be right. Therefore there exists some point between the two which separates "right" from "wrong". Just because we might not know where the centre of gravity of Dumfries is doesn’t mean it doesn’t exist.
> The survey was one on PTSD among soldiers, not on Iraqi casualties, so it may be pushing it to suggest the researchers were trying to get a high civilian casualty figure.
Not pushing it, though, to suggest that they might have been trying to get a high PTSD figure.
> anyone sensible against whom we’re fighting a war will use civilians as human shields
I wish that were the first time I’d seen the word "sensible" used as an implicit synonym for "barbaric". Unfortunately, that sort of thinking seems to have become very popular lately.
> . If we killed every single person in Iraq, then the war would be wrong. If we killed nobody, it would be right.
True. (Although the war could be wrong even if no-one was killed, but I realise that’s not what you’re on about.)
> Therefore there exists some point between the two which separates "right" from "wrong".
Untrue. A war isn’t a mathematics problem. If you want to analyse it in this manner, then there is arguably a very broad range, not a point, and that range changes constantly in reaction to a multitude of other factors. I personally think that the entire likely range more than encompasses the likely death-rate of a modern war.
I wish that were the first time I’d seen the word "sensible" used as an implicit synonym for "barbaric".
If you’re fighting a war against a far greater power and you want *you* to win *now*, then the only rational courses open to you are barbaric ones (if you want your views or your countrymen to win out in the future, barbarism is not required.)
Look no. I really am not going to have the NEJM subjected to the kind of ignorant politicised bollocks that the Lancet brought on itself. The paper was about PTSD in American soldiers. The extrapolation of 60,000 "responsibilities for noncombatant death" was my back-of-the-envelope calculation to see whether the order of magnitude was the same. Please, people let’s be careful about this and not let it take on a life of its own.
Okay, but I did state pretty clearly that I agreed with your order of magnitude. I just don’t personally think it matters.
Of course there is. If we killed every single person in Iraq, then the war would be wrong. If we killed nobody, it would be right.
No. I’d disagree with that. If we killed everybody, it could still be right. Depends on why you went to war, and how much you value human life – i.e. how you define ‘right’. This is why the humanitarian case for war is pretty ropey.
Although I’ll admit that’s a pretty ropey statement, and could be read as condoning genocide, which I don’t.
But the corollary to dsquared’s point that there is a number at which the war becomes right is that each life has an implicit finite value, which is also pretty cold and utilitarian.
Just on the subject of a ‘tipping point’, I have to say that I philosophically agree with Squander Two. Which is unusual. If we are considering the old example of the wall, painted blue at one end and green at the other, with a careful blended blue-green gradient in the middle, we have this situation: Blue at one end? Undeniable. Green at one end? Undeniable. An identifiable point at which the wall becomes green (or blue)? Not present. Logically speaking there is a point of blue/green change, but this is not determinable by humans with any consistency, so the idea of a ‘tipping point’, while logically true is, in practice, of no value.
> If you’re fighting a war against a far greater power and you want *you* to win *now*, then the only rational courses open to you are barbaric ones
No, that’s not necessarily true. It is only true if you have a certain level of contempt for human life. Also, the main reason for wearing uniform is to make war safer for your own civilians — not doing so may make life difficult for the enemy, but not half as much as it makes life dangerous and short for your own side. Abandoning that advantage is not particularly rational. It’s not necessarily irrational either: it’s a matter of weighing pros and cons. My point is that to say that it’s the only rational choice is wrong.
For most people in most societies, the rational decision-making process takes into account what sort of a people they wish to be. What you’re calling the only rational courses for anyone are, rather, rational courses for barbarians and irrational courses for civilised people.
Andrew B,
You described that far better than I did.
anyone sensible against whom we’re fighting a war will use civilians as human shields
I wish that were the first time I’d seen the word "sensible" used as an implicit synonym for "barbaric".
S2 — I think "barbaric" was adequately covered by "war."
Andrew, I think you’re fudging. I agree that the "tipping point" itself isn’t precisely determinable, but — if I understand your argument — that doesn’t mean we could simply have nuked Iraq, killing all Saddam’s henchmen, torturers, and what-have-you, along with every ‘innocent’ civilian. There’s an area, if not a point, where the cure is worse than the disease.
No, that’s not what I meant. What I mean was that we can discrimnate perfectly capably between green and blue, but that does not mean we can indentify a tipping point, and given that we can’t, we shouldn’t allow ourselves to be bound up by por-war minimisers of human destruction by being cornered into providing one.
We can perfectly reasonably say that, say, 100,000 dead is unacceptable.
There was a similar argument going on at Lenin’s Tomb, incidentally.
Dave,
No – depends on what you are fighting for. If you are fighting to liberate the people, sure. If you are fighting to prevent a WMD attack (a real, present, probable attack, not a Tony Blair one), it could arguably make sense to nuke the whole country. It wouldn’t be pretty, but it could be strategically wise, especially if the enemy has a similar level of tech to you. Not the case in Iraq, sure, but in a genuine case of us-or-them, not so unrealistic.
Andrew B,
I don’t think it would be wise for the anti-war crowd to say 100,000 is unacceptable, because the obvious riposte is ‘How many would be acceptable?’ And why is it okay to kill x many people but not y? As I said above, that effectively prices a human life, which may be empirically true, but it’s dodgy ground to argue from.
> I think "barbaric" was adequately covered by "war."
Do you? Gosh.
"If we are considering the old example of the wall, painted blue at one end and green at the other, with a careful blended blue-green gradient in the middle, we have this situation: Blue at one end? Undeniable. Green at one end? Undeniable. An identifiable point at which the wall becomes green (or blue)? Not present."
Surely though you can say ‘it’s too blue’ at some point along the wall?
Logically, in the abstract, yes. But practically the point will be indeterminate, a subjective assesment by a person at that particular time (subject to revision when considered again, say, after lunch). But the point is that the ends are indisputably ‘green’ and ‘blue’, regardless of the fact that we cannot draw a neat dividing line (a tipping point). By concentrating on the tipping point in cases such as war we allow people to deny the unacceptability of 100,000 dead, as we are pushed into providing a tipping point. But this is, in practice, an unreasonable (in the sense of it being against reason) argument cleverly disguised as a great peice of destructive analysis.
This would also mean that over 200,000 Iraqi soldiers were killed. A number given by Robert Fisk as well as some others. This also does not include the number of maimed and injured. Taking a wild assed guess, 10-20% of the Iraqi population would end up on disability directly as a consequence of this foray to steal their land, oil and water. And the sooner the US government raises taxes to pay for this the better and easier it would become.
An interesting comment left at Crooked Timber back in August last year:
http://crookedtimber.org/2004/08/10/how-many-troops-does-sadr-have-exactly/
Comment 26.
"A general point is that soldiers generally tend to overestimate enemy deaths- not just when, as John Q suggests, the war is going badly for them. (Just to be clear: the war is, I think, going very badly for the Coalition right now.)Any history of the Battle of Britain has the RAF pilots overestimating enemy ‘kills’ by 300% or more: a more accurate figure was privately arrived at by RAF intelligence who collected shot-down planes on the ground, but the higher estimates were published to aid morale. On the ground, I can think of lots of examples. Martin Middlebrook, in ‘The battle for the Malvinas’, recounts the Royal Marine patrol on (I think) Mt Harriet, who bumped an Arg unit and shot them up before getting the hell out: they reported at least 18 enemy deaths, but Middlebrook dug out the Argentine papers and witnesses and found it was three. His whole book is replete with such examples of British over-reporting: and remember, the men reporting the inflated body counts were not cowboys but some of the best light infantry in the world.(They also had little motive to exaggerate: there was no ‘body count’ policy as per Vietnam.) It’s not hard to see why squaddies exaggerate: a lot of terrified, hyped-up men all firing at individual targets, which may hit the ground because they’re dead or because they are taking cover- who knows how many get killed."
No comment on the war itself from me, just a thought on the way in which deaths get reported.
Might there be a 300% over estimation as the RAF figures would indicate? Or 600% as the Falklands’ ones? Your choice I guess.
I’m not sure if this effect would apply to this counting method: the question is effectively "have you or have you not killed >0 civilians". I’m sceptical that soldiers would inflate a body count from 0 to >0, even if they would inflate 3 to 18.