9 months on, and my challenge to engage with a scientific study remains open.
Elsewhere: Tim Lambert takes care of the innumerates, again. Barista weighs in with relevant analogies. Crikey leads with the story. WBB comments on a related matter. Tim Dunlop sums it up:
A huge numer of people have been killed and the invasion has degenerated into a disaster on any number of levels. Outrage over a peer-reviewed study by credible analysts on the basis that it âdoesnât sound right to meâ? really doesnât cut it as any sort of reponse.
Update: In comments, Leinad links to the ever erudite Kieran Healy at Crooked Timber. Definitely worth a read.
Update: Lancet cited by US Senate!
I also pointed to a Medicins Sans Frontier photographic exhibition on the Congo, which will be at the Gasworks Arts Park in Melbourne until Oct 15 2006 before travelling to Canberra, Adelaide, Sydney, Hobart and Shepparton.
Thanks for this link round-up, and thanks even more for not linking to some of the nay-sayers. The flippancy about the death toll amongst some of the people disputing this report is quite pathological.
Two instructive things about this post:
1. No mention is made of the ridiculous numbers the lancet came up with.
2. Nobody is touching it with a ten-foot pole.
It’s a links post, Yobbo.
If you want to start a discussion, go frit.
What’s your problem with the study? Results don’t fit your worldview? Or do you have some methodological problems with it?
Extrapolating 500 proven deaths to suggest there were 650,000 actual deaths is quite absurd, even if it is statistically correct. The point is that a statistical extrapolation of that kind serves no useful purpose and the entire study is an obvious political stunt.
Even Iraq Body count said the numbers were ridiculous.
Okay, so let’s see the official US stats. Their the only folks in a position to get better numbers, and they aren’t publishing them and refuting Lancet.
Why is this?
Either because they are as bad as the Lancet figures, worse than them, or better than them but still so fucking bad they can’t bear to admit to them.
Which do you think it is?
ahem, THEY’RE
Or do you think that getting good numbers on how many people are dying in a country you’ve come to ‘liberate’ isn’t worth doing?
Statistically correct but still wrong?
Translation…….my faith based belief is unshakeable in the face of scientific determination.
C’mon Yobbo, you know it’s a clusterfuck, don’t even try to argue otherwise.
What Yobbo said but in spades. If anyone believes this survey you would have to accept that on average 500 people per day have died in Iraq since the war started. Nonsense and only a dummy could believe this number.
how “amazinginly” co-incidental the last reported lancet survey came out just before the 04 election. Now they’re having a second bight at the cherry in the mid terms. How not surprising…. How absurdly unsophisticated.
All those who believe in the survey have to explian how we missed 500 people a day getting knocked off.
Phil asks:
“Statistically correct but still wrong?”
Yes Phil. The closest analogy is what once happened on Airbus planes having a faulty fly by wire system. The system was telling the pilot everything was ok, yet if they followed computer guidance without manual over ride they would be both correct and very dead.
That’s your closest analogy? It doesn’t even involve statistics ferchrissakes.
Anyway, what do you think of my last few comments JC?
It’s not perfect, but it’s the best study we’ve got.
FDB: I’m sure the actual numbers aren’t great. No doubt the entire Iraq exercise has been a disaster for the US and Iraq. That doesn’t make it the lancet figures any less ridiculous.
As I have already said, the fact that DK wasn’t even willing to put the number in this post reveals that even the left thinks it is ridiculous.
I see JC has his talking points in order. Airbus? WTF? Is that meant to be an argument?
So what are the actual numbers Yobbo? I just want to know how many deaths it takes to make you (and your soulmates) take the study seriously.
Ah, yes.
Still more RWDB special pleading.
How many people did the Stalinist regime kill? 10m? 20m? 60m?
Everyone has her own fave number.
Well, that number was generated by historian Robert Ronquest. (He later reduced his estimate to 10m from the original 20m.)
Which number do you like RWDBs?
Well here’s the scoop. The number was arrived at by the identical methodology (figuring “excess deaths”) as this Iraqi/US estimate.
Here’s a scholarly account of Conquest’s methodology:
http://freedomspeace.blogspot.com/2005/04/how-many-did-stalin-really-murder.html
So, which RWDB is going to be the first to let Stalin off the the rap that his regime was responsible for 40m deaths?
Why? What hard info are you basing this on? Iraq is to all intents a media blackhole. What the hell has your common sense or mine got to do with it?
And how much lower should the number be before we should stop screaming blue murder (pardon the pun.) 200 a day? 100 a day? And for what?
OTOH, Yobbo has at least half a point. Intended or not, this report, courtesy of its timing, is massively politicised ( and ‘common sense’ tells me this should knock the already reeling repubs to the canvas, what do y’all think? ) Anyway, I reckon we should interpret this study cautiously, be very suspicious of media reporting of it and be very cynical about any politicking of it.
That’s standard practise for all social and scientific research, let alone such a far-reaching and inherently politicised report as this one. Having said that I can’t help but mention that the report attributes approx 31 percent of all deaths to the COW troops. And we haven’t even started on the injured.
I’m afraid the numbers of russians killed off by Stalin is of little direct interest to me. (Sort of like the numbers of Chinese killed of by Mao)
But when did Lancet cease publishing articles on medical advances/discoveries & start on body counting in Iraq?
Does Lancet publish a body count for every country on earth?
Has Chimpo invaded every country on earth and said “we don’t do body counts?”
I also have my own concerns about the Lancet’s method (for starters, some zones are clearly worse than others, making extraploation problematic); but Im wondering about this more generally.
Specifcally, at what point in the rising civilian death toll would there be general agreement (nutters and extremists excluded) that it just wasnt worth it, in terms of “Iraqi freedom”? Thats the violence and complete state failure, and destruction of any real sense of security, and sheer volume of dead civilians – finally outweighed the benefits of ridding the country of Saddam (which are considerable).
Already, its quite clear considerably more Iraqis die now per day than under Saddam’s regime. And for two years now. I dont think there’s any doubt about that. But of course, all liberation struggles require sacrifice. At what point does this death rate become a net negative to the Iraqis? Have we already passed it?
Id feel differently is there was COW support for a Kurdish state, or something actually worth dying for. But with the current borders the best they can hope for is a strongarm client regime which empahsises political stability over democracy, and I fear, more or less of necessity, one ultimately as oppressive as Saddam.
Meanwhile, its a terrorist-creating magnet. Hasnt any right winger yet learned that if you take nationalism out of the equation, terrorists movements go belly up, and turn into a few hardcore “real-IRA” style loonies, with zero popular support?
Shame they’re not bright enough to catch on. Might have saved the world a lot of grief.
The whole thing’s a disaster. And now they’ve admitted it too – the line now is, effectively “yeah thats true, but how can we leave now?”. Unfortunately, there’s something to the latter phrase, now they have busted it. I dont know the answer. Its sure isnt “Iraqification” as they claimed two months ago. Now they’re blaming the Iraqi security forces for much of it! Shit. What a complete neo-con wreck.
The answer to this question may be rolled under the beds of the more gung ho officers wrapped around a frag grenade with the pin removed.
FDB, Phil
For those numerate inmunerates a statistical method can spew out what an average IQ with a little commone sense would see as ridiculous.
Can either of you demonstrate a day in Iraq with 500 deaths being recorded.
Phil, FDB the example was to show that you cannot always rely on proscess to achieve a correct answer when common sense and inituition mark it as clearly improbable. You two obviously think that 500 deaths per day is about right. Is that what you are actually inferring?
Michael G
I didn’t know that post was about the pros and cons of the Iraq war. I thought it was a discussion on the merit of Lancet and it’s validity.
LE says:
“Id feel differently is there was COW support for a Kurdish state, or something actually worth dying for”.
So in LE’s world its ok to die for Kurdistan and say Dalfur ( I presume you would agree with the last one) but not an entire nation. So LE has deconstructed intervention only to regionalized conflicts within a country and not a nation itself. Now that’s an idea!
http://www.atimes.com/atimes/Front_Page/HI14Aa01.html
It is well known that many bodies never make it to the morgue, it being too dangerous for Sunni to turn up at some morgues and Shiite to turn up at other morgues.
However, this article provides a useful metric. Baghdad authorities want to process 250 corpses a day. Fewer than 1/3 of Americans die in Baghdad. Many more die in el Ambar Province. So, much fighting takes place beyond Baghdad.
Therefore the 250 per day projected throughput in Baghdad could easily add up to 500 per day nationwide.
Don’t forget that only those who die by violence or misadventure ever end up in a morgue.
Few of us reading this post are likely ever to find ourselves in a morgue. That’s one of the rarely considered benefits ofnot being invaded by the US.
Oh come on Joe. Another rule of thumb for interpreting research. Understand the context. No survey is conducted in isolation. This goes double for interpreting reactions to the survey. And it goes triple for interpreting reactions from those of a hawkish persuasion. And this means you. Don’t play coy now.
However, if you’re implicitly admitting that the amount of people in Iraq who are dead now – and who in all likelihood wouldn’t have been dead had the COW never invaded, or had they gotten out before the civil chaos descended – is shocky and distressing, and that we should sit down and work out who is causing what, and how we can best minimise further casualties without compromising wider security issues, well, then we can get back to more fruitful discussions.
If not, then go back and reread Tigtog’s first comment.
Its open to the US/UK to invade Sudan and come to the rescue, JC, on the Iraq model. But they wont. No strategic value. So get off that high horse, cos its a hogtail nag. Its just a UN-bashing exercise. A UN desperately short of blue helmeted troops to fill existing PK operations: to which the US and UK contribute essentially no-one. But, boy, can they whinge about the UN, and demand new peacekeeping ops for it? Sure can!
I find nothing to respect in their position.
And JC, lots of Iraqis are dying for “their nation” as we speak. And they’re trying to take out as many COW soliders as they can, when they do. See the problem?
Michael
I’m not debating whether Iraq was a good gamble. i’m talking about the efficacy of the survey. Clearly reality is out of whack with lancet. Nothing wrong with Lancet per se. As I am not a statician I can’t venture to say why it the method used is incorrect. But it is unless you believe 500 people die daily in Iraq. Do you?
It extremely unlikely there are so many dead.
Katz says
“Therefore the 250 per day projected throughput in Baghdad could easily add up to 500 per day nationwide.”
You are a propagandist and no one should take you seriously about such issues. It is not up to you to speculate, it is up to you to show us proof. How many days are recorded of 500 or more people dying per day? Details please.
Ah yes, the Windschuttle dodge.
On that basis Stalin killed no one.
katz
your last post made no sense. would you mind explaining please. Also you may want to prove the daily count exercise I left with you.
Katz
The war has been going on a for a few years now. So there must be a record of that many killed in one day. Show us how many days there are with those numbers. As I said you are a propagandist and required to verify nearly evrything you ever say.
err Yobbo and JC, extrapolating to a population the information from a sample survey is what statistics is all about.The ABS do it all the time.
I do note Zogby has commended the way in which they did their sample survey so I suspect it was done the way most sample surveys are done ( think newspoll or neilsen for example.)
Quite a few Bush and Howard apologists should read up on statistical technique.
And yet you accuse others of being propagandists? Get real JC.
Here is why I believe the number. It’s called a peer reviewed process, and it’s used regularly in the medical field and has been used for decades in research that has contributed to our present well being in that field, I think that means it works.
On the balance of probability that the study represents I’ll accept the number as being correct.
And don’t ask me to how many days. because you know as well as I do that’s not how it works, that’s almost as bad as your Airbus analogy.
Homer
I then ask you to comment on the question i asked earlier. Go for it!
it is late at night JC and I am going to do something else soon.
Which question is it?
I’ll try to put it in short words that you may understand.
There are few official records of deaths during Stalin’s regime.
Do I have to tell you why?
That being the case, researchers may conclude that there were few deaths.
Now we know that that’s ridiculous. How? By indirect methods pioneered by Robert Conquest. See my post on this above.
The Lancet study uses similar methods, because Iraq’s official records are incomplete.
To make the matter quite clear, the Iraqi records are probably not as incomplete as the Soviet records. But one would get quite a misleading picture of both Stalin’s and iraq’s deathrates were one to use only official records.
JC. You know and I know, and quite a lot of people around here know, that you wouldn’t be worked up about this if you didn’t already have strong views that predispose you to apologise for whats left of the Coalition of the Willing.
If this thread turns into a straight discussion of the methodology of the Lancet – and even there, you’re common sense line ain’t cutting it by the way – then it is missing the far greater point.
If the study is anywhere near the truth then it is a scandal. One that some of us are not that surprised about as you, but a scandal nonetheless. Dk’s original post was fairly open and so i think comments about the priorities of people attacking this report are as on-topic as anything. As Tim Dunlop says:
Phil
You couldn’t understand the Airbus analogy because it may be beyond you and for that I apologize. It was easy: in fact it explained how people should not always rely on process (machines in the Airbus example) when common sense judgment and intuition are clearly superior in figuring things out.
“On the balance of probability that the study represents I’ll accept the number as being correct.”
What’s the basis of assessing the probability… the methodology itself? That’s circular reasoning and a dog chasing its tail would appear better able to cope.
The best way to figure out the veracity of the number is to look for evidence that it has some “probability” it is pointing things in the right direction. A way is the see if there are any days on which that number hit the morgues. I can’t recall one day I have ever seen after the fall of the capital when that number has reached ceiling of 500. And yet you calmly accept that because process like in the example I presented (the plane’s mechanics were signaling all was fine) was ok you would follow the instruments and not use your judgment and intuitive skills as a pilot. Captain Phil, you just let the plane crash without even bothering to notice the nose is pointing down.
On the methodology.
This study was conducted using tried and true epidemiological survey methods, methods currently in use by the US Govt. for their own disaster relief projects.
The same methods were used to calculate the deathtoll of the Congo War (The other great massacre of the early 21st century), producing figures of 1.7m dead, figures that Tony Blair and others have had no qualms with using in their speeches on that topic.
Similarly, cluster sampling methods have been used to calculate the loss of life in the Darfur catastrophe, where they derived a figure of 600,000 dead, in over two years of sectarian carnage, mass displacement and the complete disruption of services and breakdown of law and order (sound familiar?). Do the Lancet study’s detractors deny there is a mass humanitarian crisis in Darfur as well?
This study was peer-reviewed and picked to pieces by three other leading epidemiologists with no connection to the study’s authors.
They found no problem with it.
Put simply, if you quibble with this study’s methodology you’re obliged to think that either:
a) the entire fields of statistical analysis, epidemological research and disaster relief are bunk,
or
b) all epidemiologists are in a conspiracy to destroy the Bush administration.
Take your pick.
I am going now for the night.
My understanding is that there was a random sample made of Iraqis from whom the researchers gained their information.
I believe it would have been more accurate to say a range was involved given the margin of error as should have been in a previous survey.
I do note that neither Bush nor Howard have stated their objections to the statistical methods used.
As I previously Zogby has given it the big thumbs up and he does know something about statistics.
Eh. I was a bit overzealous. Replace “quibble with” with “reject”.
Well, the Iraqi Ministry of Health itself estimates civilian dead at 128,000. And I think we can all well imagine the difficulties they might have in collecting a full body count.
But it would, presumably, cover those who died in hospital or ending up in morgues, or some state apparatus where counts would be routinised.
That in itself suggests to me the figure has to be at least 200,000; and that would be very conservative.
So, regardless of your views on Lancet’s methodology – the invasion and its wake are already occasioned a fully fledged humanitarian disaster, of Tsunami magnitude.
Incidentally, what “methodology” do you think GW bases his best guess of 30,000 dead? It sure doesnt involve talking to those doing the medicial treatment and burying (see above). This gets the prize for the least credible count.
Leinad
You left out C, which is the correct answer.
c)The incorrect application of methodology used to calculate death rates in circumstances that present a unique set of criteria that we yet don’t understand.
Happens all the time.
You don’t seem to be getting it either.
The death rate estimate is far too high for it to be accurate.
It means we need to go back to the drawing board and figure out why the method is not correlating to the real situation. Could be a Phd puzzle for a really smart math wiz. maybe you ought to try
Guys i don’t know why you’re faliling around trying to concoct all kinds of evidence based and scientific arguments, when we all know the truth of the matter.
As JC says, the death rate estimate is far too high to be accurate. It just is.
Maybe JC is getting freaked out by the sheer size of the number. I was at first.
To put it in context, JC, 640,000 is 2.5% of the Iraqi population. That’s one in forty people.
So, to believe that, you really only have to imagine that for every 39 Iraqis, selected at random, one of them had lost a family member in the conflict of the last three and a half years.
And now that Ive put it that way, to myself – suddenly I dont find the Lancet figure so hard to believe at all.
“Well, the Iraqi Ministry of Health itself estimates civilian dead at 128,000″
Not in this article LeftyE. I have no real idea how many have died overall, but if we accept a figure around 100,000, it does seem unlikely that you could hide another 500,000 bodies.
I’ll wait and see if the figure is discredited
PIMF – “if we accept an Iraqi figure around”
LE asks
“Incidentally, what âmethodologyâ? do you think GW bases his best guess of 30,000 dead?”
The Pentagon as I recall.
It’s interesting how you guys are against the war but cheering for a higher count. You’re such ghouls deep down aren’t you.
Looks like JC just surrendered.
Not that I blame him. It was unsustainable.
Anyway, please highlight any “cheering” above JC (as I cant see any); or just admit that was a dummy spit, and we’ll get on with the thread in an amicable fashion.
Ok, I’ve added a third option.
C) Joe Cambria’s gut tells him the number is too high.
In which case, Joe, could you hold your ear as close as you can to your stomach and find out the correct answer. Take your time, Johns Hopkins University is more concerned with accuracy than speed.
Oh and:
is a sufficent encapsulation of Joe’s grasp of the issues.
Bush’s number of 30,000 was the lowball estimate provided by (the horribly evil and leftist except when it’s making us look less gore-soaked) Iraq Body Count in Dec. 2005.
The Pentagon, famously, doesn’t do body counts.
LE
I’ve known you too long not to be amicable, though at times like this you are trying.
LE. The number is bullshit and it’s meant as a political device to hit the GOP in the mid terms. Like the previous study, it always seems to make an appearance just before US elections. It’s almost like the coffin lid opens after dusk with these guys.
Can I statistically prove it’s wrong? Of course I can’t. Can I intuit that it isn’t right? Sure I can because it is magnitudes over other valid ways of looking at the numbers.
Here’s an offer. I’ll offer $2,000 bucks to anyone who can show me 10 consecutives days that the death toll has been over 500 since the fall of Baghdad. How long has the war being going on for 3.5 years? With that death toll there would have to be 10 consecutive days of 500 or more over this period. Show me the evidence and you get 2g. Start googling LE.
And I have always had a liking for you Lefty so don’t suggest otherwise..
Leinad
I never said I believed the Pentagon, did I? So please read and discuss what i say with some degree of accuracy, otherwise don’t bother commenting.
I don’t know if the Pentagon is right or not. I certianly wouldn’t even suggest this here where I have read comments of some who think GW was the cause of 911.
I wouldn’t for the minute attempt to convince you of the veracity of Pentagons numbers.
Just a footnote for PeterTB: I saw that Iraqi Ministry of Health figure (128,000) on SBS news, earlier tonight.
I guarantee you that’s the figure that came up on the screen. As for their source, you’ll have to ask SBS.
Yobbo says:
“Extrapolating 500 proven deaths to suggest there were 650,000 actual deaths is quite absurd, even if it is statistically correct”
What a remarkably moronic comment. You are effectively saying that fields of enquiry that rely on statistics, like epidemiology, are completely invalid. By golly you are brainless.
He’s got you there, LE.
Unless you can scour the morgues, crematoria, necropoli, collate the death certificates and eyewitness reports, unmask the unmarked graves and dredge the Tigris and lesser waterways of a nation of 27 million (in the midst of a civil war no less) for the required number of corpses I will have no recourse but to declare Joe’s gut the winner.
Heaven help us all.
Not hard to find sites with death toll figures which are totally inconsistent with Lancet though. They’re everywhere
LeftyE – I wasn’t suggesting any dishonesty on you’re part. Sorry if I gave that impression.
Leinad
you are so insincere and always in bad faith. Never stand between leinad and a dishonest side of an arguement as he wins all the time.
If we had that death toll, there would be a record of daily death counts recording 500 or more. You don’t have travel to Iraq for the numbers, a good google serch would find it.
So don’t talk rubbish and please keep it honest for once in your life.
argument ….sorry Steve.
Well, indeed. Because we all know what Joe’s (and everybody else’s) gut is full of, do we not?
Steve M
Yes, statistical sampling works well most often. In this case it does seem to have broken down. it is either becasue the survey takers were totally dishonest and cribbed the stats or there is something wrong with the methodology used in this field.
We don’t know and Lancet is certainly not going to tell us, as they would be the best sources.
However statisical sampling by using 600/ 27 million is a bit light on i would think. lets not forget that election surveys use a bigger sample in Australia and have geographical issues etc figured out but they still get thigs wrong at times. Iraq is nothing like that. It’s a minefield – pun intended.
No worries – I didnt take it that way, PeterTB – just sourcing my previous comment, since you’d found a different one elsewhere.
JC, I hope you wont think me a coward if say: I wouldnt set foot in Iraq on some data collection mission for quids.
I understand this was also a key challenge for Lancet’s researchers. But that hardly bodes well for someone who wants to poke holes in the method: that’s likely to mean they didnt go to the most dangerous areas while sampling.
And as for surfing webpages – I mean, isnt the media’s problem precislely the same? Too dangerous to report?
I cant see what that will establish.
Better than your head Pav. At least mine’s in the right place.
LE
“JC, I hope you wont think me a coward if say: I wouldnt set foot in Iraq on some data collection mission for quids. ”
In this case LE I’m a gentleman and allow others to walk throught that door.
Awww.
More theatrics from our flailing gut-thinker.
Joe, you don’t seem to be able to grasp the fact that cluster sampling and extrapolation are valid statistical techniques (they’re the mainstay of epidemiological research, particularly in disaster zones or areas with heavy civil conflict like Darfur and Iraq).
You don’t seem to know where Bush got his 30,000 figure (hint: not the pentagon), nor do you seem to understand that there’s no way every death in the entire country of Iraq is going to be reported, especially in the middle a civil war where kidnappings, executions and mass graves are rife.
Rather than actually challenge the Lancet studies on methodological grounds, or, failing that, provide a peer-reviewed study of your own that challenges the Lancet figures, you’ve resorted to argument from incredulity, casting nasty aspersions on other posters motives and petulant sulking.
Oh and as far as bad faith goes, you can’t beat this:
Dishonest indeed.
PS: This is a gem,
Joe Cambria:
Leinad, this one is my personal favourite piece of Joh Speak:
“You don’t have travel to Iraq for the numbers, a good google serch would find it.”
Well goodness, gracious me.
No, no, I insist, you first!
Do report back on how safe and unlikely to be killed one is in these days in sunny, post-liberation Baghdad.
And no cheating in the green zone, JC.
“argument from incredulity”
That’s another way of saying healthy scepticism isn’t it? Aren’t you the least bit curious as to why the Lancet figure seem totally out of whack with all other reports?
Joe, you donât seem to be able to grasp the fact that cluster sampling and extrapolation are valid statistical techniques (theyâre the mainstay of epidemiological research, particularly in disaster zones or areas with heavy civil conflict like Darfur and Iraq).
Do I need to repeat answer c) again? Yes I do it seems:
The incorrect application of methodology used to calculate death rates in circumstances that present a unique set of criteria that we yet donât understand.
In case you didn’t know the field of statistics is not stationary and actually does evolve over time. Methodology or theoretical applications like game theory came into existence in the 80′s I think.
What you are suggesting is that there all knowns are known in the field of stats, when in reality it is still a field with lots of unkown unkowns and known unkowns .
Black Scholes was developed as a pricing model for options. It wasnât known until a little later that it become obvious the pricing curve wasnât flat as it was assumed in the model and only later was it corrected.
According to you Leinad, banks should have ignored the real losses suffered by utilizing an incorrect model and ploughed aheadâ¦.. ignored the real hard evidence or the obvious because it wasnât specified in the model.
âBut the model was correct cried!â? Leinad, as he was escorted out of the room.
Leinad. There hasnât even been 5 days of inconsecutive 500 death tolls and you are still arguing what is obvious.
wn unkowns.
Steve M
That comment was referring to the challenge in finding 10 consecutive days where the death toll was 500 or above.
How about you reporting back 10 days in the whole 3.5 years with a 500 a day death toll. I would be interested in seeing what you find.
I’m certainly curious Peter.
So let’s see. Your link says 3400 in one month (the highest monthly count to date, apparently.)
Adjusting for it being an excessively bloody month, we get about 90-100 a day. Or about 100 000- 130 000 overall, no? Awful close to LE’s figure. Because I’m tired and lazy, i just took the 500 a day equals 650 000 figureand extrapolated. So 100 a day = 650 000 divided by 5. Feel free to correct my maths.
Oh, ive just seen your commondreams link. This seems to offer a stronger case. A straight up 50000 count, and a bunch of Iraqi sources (Baghdad Morgue, Ministry of Health etc.) But further down, there is this:
As an aside, this is also interesting:
Peter:
Argument from incredulity is when someone says something like: “The lancet stuffy is iffy, the numbers are too big for me to believe it”.
Skepticism means coming to a position through careful study and evaluation of the evidence. Argument from incredulity chucks out the evidence because it doesn’t feel right.
The Lancet studies get different figures because they have different methodologies. Iraq Body Count, for example, collects and collates media reports of casualties in Iraq to derive a basic ‘floor’ for civillian and military casualties. The Lancet studies used cluster sampling of 1,800 Iraqi households chosen from 47 ‘clusters’ across Iraq to determine the death rates for pre- and post- Invasion Iraq and compared the two.
Both results are different, but IBC can only mark the deaths that get reported in teh media, so isn’t anywhere near a comprehensive figure.
To back up Leinad’s point. From the Lancet report:
So if the 20 percent is approximately 100 000 – 130 000, where does that get us?
They have a list of factors which could cause error and bias, some of which would cause over-estimation and some underestimation. And here’s there finally reccomendation/plea:
damn pdf’s. Sorry for the ugly blockquoting
And now, straight from Joe Cambria’s gut:
Leinad says:
” You have yet to get your head around the idea that not every death, every mass execution, kidnapping, or petty murder in Iraq is recorded in the media … ”
Good point. And I also suspect the number of deaths that get reported may be higher in the Arabic language media. I wonder if Joey could do a google in Arabic?
Leinad
The lancet numbers is a huge outlier that no teven you can justify without bluster and bad faith debating skills which I have to admit you’re proving to be a master.
The first lancet came out in 04 i think right…. just before the prez election. The reported 100,000 deaths then. So since then we have gone into acceleration mode and in they are suggesting ther death count is now approx. 625,000 – 100,000 /2/365 = 720 deaths per day.
Are you rational? Do you actually believe there have been 719 deaths per day since the last survey?
You wanna set up a bet on this? Figure out what you wanaa do, how we escrow and lets put a little money on this.
This is my last one for the night, I promise
.
Calling the Lancet number an ‘outlier’ pretty well sums up your numeracy, jc.
The different measures like passive reporting and cluster sampling aren’t a single population.
It’s been a long day out in the real world. Thanks to Leinad and Michael G for filling in the blanks here and there and Yobbo for starting this 80-odd comment thread by saying there’s nothing worth turning up for. Perhaps that’s right, assuming you followed the last faux-debate. I’m off to bed.
I know jack squat about statistics or scientific methodology, so I won’t address that. I merely offer some other generally accepted death counts in large recent historical conflicts; you can use them for purposes of analogy, if you like.
Iran-Iraq War: lasted eight years in the 1980s. A clash of two massive conventional armies, with regular campaigns and regular shelling, the long-term siege of cities, minefields, the works. (as well as the use of chemical weapons on a large scale.) Total casualties are I believe somewhere in the neighborhood of a little north of one million: that is, about 45% higher than the 625,000 claimed for a conflict of half the duration, in which only one of the many parties involved is fully mobilized and massively armed (and said party mostly does not use anything close to its actual military capability).
France in World War I: out of a total population of 40 million, France experienced casualties of 1.3 million; and this in what is considered one of the bloodiest wars in history, again with massively mobilized armies, chemical weapons, modern killing machinery, no-man’s lands, etc. Roughly twice the estimate of 625,000, for a population about 40% larger than Iraq’s, incurred in a conflict of such great size that it was impossible not to notice it if you were alive on planet earth at the time. Compare this to a 4-year civil war in which only one side is mobilized into regular units, and in which most of the killing is low-tech, locally based, not organized in campaigns, and done mainly in a clandestine manner.
US Civil War: lasted four years, death toll roughly equal to the number claimed in Iraq by the Lancet. In order to come up with the same number of deaths which the Lancet claims for Iraq, a nation half the size of a continent, with a very large population, had to split itself into two gigantic armies which fought with artillery, massed fixed-bayonet charges, and campaigns of massive civil destruction for the same length of time as the current conflict, in which, as I note, the methods of fighting are mostly not comparable in their ferocity, organization, or techical advancement.
On the other hand, the killing in Rwanda seems to have been accomplished in record time, with fairly crude tools and methods, so perhaps that counts in the Lancet’s favor. But then, it may have had other characteristics that distinguished it, as well.
What does it all mean? Damned if I know. I’d like to say I believe in scientific methodology, but I can’t help thinking JC’s got a point. The truth, as usual, is probably some place in the middle; in the meantime, of course, there’s absolutely nothing to be happy about, even if the numbers were far smaller than claimed.
As Mr. Plant once said, “And it makes me wonder.”
Leinad.
I never made any representation that I studied statiscal analysis. So stop playing the strawman.
âWrong. Joe wouldnât know a 95% confidence interval if it kidnapped him off the street, shot him in the head and buried him in the desert.â?
Well actually I do. And here you go again with a bad faith debating style.
âPots, kettles, hell chuck the frying-pan and the wok in – this is guy who accuses LE of wanting more deaths in Iraq talking about bad faith.â?
And proving it again in this sentence.
â98,000 was the most likely figure, with a 95% confidence interval of 8000-198,000.â?
Yes. Knew that, thatâs why I used 100,000 from the 04 figure.
âIf Joe knew what a bell-curve distribution was, and how it worked, heâd know that the central figure of 98,000 excess deaths was the most likey figure, and that the higher and lower ends were progressively more unlikelyâ?
So I use the figure lancet 04 used (100,000) â¦. I get a lambasting and a sermon about a bell curve. Letâs note here that you donât seem to have a problem with the 04-survey prediction. So this sentence is useless and another bad faith attempt.
âThis is a classic sign that Cambria hasnât read either study.â?
Which I did, but thatâs you lying that I didnât.
âThe latest Lancet report is an update on the figures of the 2004, bringing them up to the end of June, 2006.â?
Yes so what. I never said otherwise Are you making a point here or simply trying to divert attention from god knows what this time.
âNot a report of 655,000 further deaths. As for âacceleration modeâ what part of wipespread sectarian carnageâdonât you understand, Joe?â?
Thatâs not what I said, Lenny. What I did was reduce the overall total since 04 by 100,000 arriving at 525,000 (it should have been 555,000 as you say). This increases the estimate from my previous fig by25,000 pushing the extra death toll up to 760 extra death per day as a result of the war. Summary: Lancet is telling us there have been 760 extra death per day since 04.
âIâm rational enough to read a study before I critique itâ?.
You were obviously sucked in without even thinking about it.
âThe fact that youâre waving 719 a day around shows that you havenât made the slightest inquiry into these studies that you âknowâ are incorrect.â?
Its now 760 extra deaths per day as you corrected the number by 25,000 since 04. I didnât say it was incorrect Lenny. You assume Iâm saying that or acting in bad faith again. I have maintained the methodology looks like it doesnât work well for Iraq for reasons we yet donât understand. Thatâs because other evidence shows it is out of whack.
And you still havenât presented even one day when there was an independent report that the death toll even approached this horrifying number. NOT EVEN ONE.
âTo place bets, you actually have to know what youâre betting on.â?
Easy. Clear. Bet there have been 760 deaths in 10 consecutive days that are verifiable by other sources. In all of the 1277 days of the war (3.5) years
âJoe doesnât know a passive survey from a cluster analysis.â?
I have a working knowledge of the process. You donât by the looks of things otherwise you wouldnât be reaching to the extraordinary height of attempting to prove 760 people die a day as a result of the war since 04.
Iâd rather bet with Sol Trujillo.
You sure would. You just bought Lancet bridge Ltd.
Lenny seeing weâre on a full name basis why not tell what your real name is instead of hiding behind Leinad and not affording other people the same dues. Pathetic!
Yes, 30,000 dead Iraqis is a much more acceptable figure than 650,000.
Particularly considering the War On Iraq was sold as a war that would inflict hardly any civilian casualties at all, and be over in a few months, if not weeks. Remember Rumsfeld and all his talk about the high-tech, precision targetting munitions?
It’s been long accepted in military circles that at least three or four thousand Iraqis were killed in the opening days of the war. Remember ‘Shock And Awe’?
Another 3000 to 6000 were killed in Fallujah in 2004.
There’s almost ten thousnad dead right there.
Face it, before this is over the death toll will be more than a million, whichever way you want to cut it.
If you supported the war, live with the toll and stop whining.
It’s pathetic.
These people did nothing to us, they did not threaten us, they did not attack us, they did not wish us harm, but tens of thousands of them (if not hundreds of thousands) are now dead, more than half of whom were children.
More than 80% of Iraqis want us to get the fuck out of their country. More than 60% support attacks on coalition troops.
The Iraqis lived up their part of the war by not fighting back for months after we invaded. We failed them by not leaving.
And JC, that is one ultra-paranoid conspiracy theory you’ve got going there about these reports coming out before key US elections. Did you get that off Bill O’Reilly’s Talking Points?
Grow up.
This death toll denial stuff smacks of those foul creatures who rant that six million Jews could not have possibly been killed in such tiny gas chambers.
There is no acceptable death toll of civilians, that is war propaganda. Our leaders will lie to us and lie to their own soldiers if they think it will help them get whatever it is they want out of Iraq.
At least the dementoid on Tim Blair’s blog who reacted to this study by saying : ‘Thin The Herd’ was being honest. We can only wish our leaders would tell us the harsh truth about all those reports they read and then bury.
Democracy in Iraq was a beautiful dream, and they have it now, for what it’s worth, but something else happened in Iraq while all that democracy was flowering.
It’s called ‘Depopulation’.
And depopulation is what virtually every major war of the 20th century has been about.
Thin the herd, indeed.
Michael G – no problem with your maths, we just need to bear in mind that the Lancet figure (650k) needs to be compared not only with IBC (say 50k) but also with the “passive surveillance” technique used by the Iraqi Ministry of Health (sometimes called counting) that yields 50k thru 128k depending on reporter.
Now in Australia, passive surveillance techniques would give us a pretty accurate estimate of overall deaths, while in somewhere like Rwanda or the Congo, I think you would get a far more accurate figure form deaths reported in the media. Where does Iraq fit in this scale? Could it be that only 1 in 5 (or maybe 1 in 10) deaths are being (have been) counted by the Iraqi MoH?
Dunno really. Seems unlikely, but probably best to keep an open mind at this stage.
Darryl
I’m not surewhy you even posted this comment. I haven’t said here that I supported the war. Moreover it has nothing to do with the fact that Lancet is incorrect.
Worthy thoughts re France j_p_z.
But don’t forget that those figures are referring only to military casualties, whereas the Lancet study is about all casualties, including the controversial component “excess deaths”, i.e., an estimate of deaths above some hypothetical background deathrate arising from Iraq being a normal, peaceful nation.
Moreover, read my above link to a discussion of Robert Conquest. More recent researchers believe that even he underestimated the number of “excess deaths” under the Stalin regime. (One figure is 60m.)
Of course, apologists want to minimise deaths. Critics want to maximise deaths. Neither side is particularly keen to nominate an “acceptable” deathrate, which is much more important than arriving at a raw number.
Exactly, Katz – which is where I started my comments last night. I ask again
My own opinion is that point was passed about a year ago. The invasion and its wake have been a net disaster for Iraqis, on any criteria.
Peter,
The excerpt I posted from the Lancet study (11:57 pm) claims that:
So that means in all probability less than 1/5 is counted. But yeah, an open mind is definitely a good idea.
AFAIK (my momma’s an epidemiology PhD and we talk about this shit all the time) cluster sampling actually tends to underestimate, when the incidence of what’s being measured is likely to be concentrated in small spots. You’re more likely to miss those spots than hit them. Then again, hitting hot spots by chance with say 20% of your clusters will bias your figures upwards.
In this case, they have avoided the most dangerous areas for researchers (for their personal safety) and the areas known/thought to have particularly high body counts in the past (e.g. Fallujah). Having excluded these areas, the samples were selected from a GPS grid by selecting coordinates randomly, then finding the nearest households still standing and occupied.
This suggests to me that they made a deliberate effort to avoid biasing their sample with hot spots, then went as random as they could for the rest of the “country”.
Add to this that they couldn’t survey households which had already fled as refugees – to my mind these families are especially likely to have suffered a death and said “fuck this, we’re outta here”.
I agree that the figure seems very high, but what the fuck do I know? Answer: what I’m told by the US military and the media of my own complicit nation.
It’s also worth noting (as many here have) where past figures have come from. Other recent studies using similar methodology have generated death tolls that nobody has been especially keen to question. Why the panty-bunching now, skeptics?
Interesting issue LE.
I guess one has to assume that the Bush Administration genuinely believed before invasion that their occupation and reconstruction strategy was the optimal one. (I’m not talking here about intelligence or insight, I’m talking about sincerity.)
Having set their policy, then presumably the Bush Administration would have set some benchmarks to measure progress. If those benchmarks aren’t reached, then the only reasonable options are:
1. Change the goal.
2. Change the methods used to achieve the original goal.
3. Leave.
Did the Bush Clique do any of these tings?
1. Change the goal.
The Bush Clique got hijacked by Sistani. US plans for a dependency collapsed under the weight of Shiite power. The Bush Clique bowed to this but never really accepted it as a desirable goal. Instead, they stayed in the game with their 150,000 troops, hoping against hope that something would turn up. It didn’t. Instead of a reasonably stable Shiite theocracy, Iraq is now anarchy. It’s a lose/lose situation for both the Shiites and the Bush Clique. I disagree somewhat with LE. I believe this began happening about two years ago. Unfortunately, the Bush Clique is in denial about this and is therefore unwilling to acknowledge the reality that every death has been an excess death for two years.
2. Change methods.
Puh-lease!! How many times can the US troops invade Baghdad? The Bush Clique has no other answer but evermore surreal military adventures. US troops are trapped in Rummy’s nightmare. The Bush Clique is in denial about this and is therefore unwilling to acknowledge the reality that every death now is an excess death. And their dying not for Iraq, but for the electoral survival of the Republican Party. What a waste!
3. Leave.
More will die, but they would have anyway. A quick victory by the Shiite forces will at least hasten this sorry story to its inevitable conclusion.
Yes, Katz, and I tend to agree with your earlier comment that fragging (or a change of administration) may prompt the withdrawal.
I’d like to see some historical data on fragging. Im guessing its difficult to collate, but Ive heard of the phenomena in relation to WW1 and Vietnam – but never WW2. I wonder, therefore, if it could be considered a gross measure of how stupid and pointless a war is, from the footsoldier perspective – or, does it also require the element of conscription?
According to Wiki (and I say that as a rider), four US officers have met this fate in the Iraq war. I assume they mean the most recent one. http://en.wikipedia.org/wiki/Frag_%28military%29
Hmmm, that link didnt work. Oh well.
And on methods, ‘Iraqification’ seems to be yesterdays news. If you gaze across the history of rebellions, insurgencies, civil wars, and revolutions and nationalist struggles, one truism emerges: Once the security forces switch side, in any significant number, the gig is up for the regime.
I agree with JC that “common sense judgment and intuition are clearly superior in figuring things out”.
Applying his techniques, I’ve figured out that JC has a one of those “female brains”. becuse he can feel things in his waters.
Welcome to the sisterhood, Joe.
Arguing about the methodology used in the paper is fine.
Arguing with reference to common sense is not.
And on a closely related issue:
Britain’s top army commander said British troops in Iraq should be withdrawn soon because their presence was exacerbating security problems in the country, according to a British newspaper.
http://news.yahoo.com/s/nm/20061013/ts_nm/iraq_dc;_ylt=A9G_RwrO4y5FeVMBBgJZ.3QA;_ylu=X3oDMTA5aHJvMDdwBHNlYwN5bmNhdA–
As in a landslide, first a few pebbles fall, then the whole mountain collapses.
Eat that, RWDBs.
Aarrg. Late nights Leinad = not so good posting.
Ok, firstly, Joe, i’d like to apologise to you for mischaracterising your argument in my last post, and for the lecture. Mea culpa. I did not intend to misrepresent you, but I was arguing out of anger rather than with a cool head.
That said, you have been overly dismissive and less than civil to several other posters in the course of this thread, IMO without good cause, and while this does not excuse my tone and my language, I think you should reconsider some of your words and the tone you have used in this discussion.
Back to the arguments:
Throughout this thread, Joe has maintained that the Lancet numbers are too big. He has argued that the proof of this lies in the fact that the daily reported civilian deathtoll has never reached 500, let alone 700 in the period between 2004-2006. This, he claims is proof that the proven methods of the Lancet study have hit some unknown factor, or series of factors in Iraq that has thrown their accuracy off, despite numerous safeguards.
Consequentially, he has challenged those who maintain the accuracy of the Lancet studies to find a single day where reported civilian death-tolls beat, or even reached 500 (or is it 700 now?) – inability to do so, he maintains, is proof of the Lancet study’s innacuracy.
There are several problems with this argument, as many posters on this board have already pointed out.
The first is that the reporting of Iraqi deaths is by no means complete. Our figures on daily death toll come from the Baghdad morgue, media reports collated by Iraq Body Count and what we’ve managed to glean from the Iraqi Interior Ministry, the US DoD(they count civilian casualties reported in the course of events to which the Coalition responded).
These organisations conduct what’s known as passive reporting – they collect data brought to them; bodies in the case of the Morgue, and media reports in IBC’s case. By definition they don’t provide the full picture, as every violent death isn’t necessarily reported, let alone nonviolent deaths. The Lancet studies work by cluster sampling – determining death rates from a representative sample of the population and extrapolating probable figures from them.
It is impossible to obtain complete figures from passive reporting, especially in Iraq. For one, sectarian conflict has devolved to kidnappings, mass executions and burials, with some bodies not being uncovered until days after the event. Corpses have been dredged up by fishermen on the Tigris, and stuffed in barrels in the desert and they are almost certainly not alone. As well, Islamic custom demands the dead be buried within a day of their passing, which reduces the likelyhood of them passing through a morgue or obtaining a death certificate.
As a general rule, the chaos inherent in a conflict situation plays havoc with record keeping and institutional memory. With the exception of the Bosnia conflict, no passive surveillance in a conflict situation has recorded more than 20% of the deaths established by population studies. This is precisely the reason why the Lancet studies were conducted in the first place – to establish via extrapolation what cannot be estimated due to the disruption of normal bureaucratic record-keeping.
In spite of this, passive surveillance is reporting a steady increase in the civilian death toll, matching the accelleration of death rate, both violent and nonviolent described in the Lancet study.
Over at Crooked Timber, Kieran Healy has a nice companion piece for people who boggle at the huge numbers presented in the report, which I think deals quite nicely with the scale issues.
In some senses, it was a mistake for the Lancet’s editorial board to emphasise the extrapolated numbers, for they aren’t the crux of the study’s findings. Validly derived or not, they’re spongy and inherently dodgy-sounding, and their magnitude alone can drive even diehard Bush critics to doubt. As FDB has pointed out, cluster sampling has bias towards under-reporting, which suggests these gigantic figures might be even larger.
However, in my opinion, the real bombshell in the Lancet studies are the death rates.
The rise in violent rates from 0.1% to 7.2% over 2002-2006 and soaring total mortality rates, 19.8% in 2005-2006, outline the increase in everyday risk and privation that Iraqis are subject to – be it typhoid from the sewage piles of Sadr City or the nightly terror of kidnapping by masked men in uniforms, killers who may or may not be your local cops by day.
These indicators of a massive upswing in violent death rates post-Invasion have been confirmed by death certificates and matched by a range of passive reporters from NGOs to the DoD. Furthermore, they have been assiduously determined by a brave team of dedicated professionals who drove the length and breadth of a strife-torn country to ask painful, personal, and above all, dangerous questions of greiving families.
Wether you think the real numbers are 30,000 or 3 million, there is no denying the lengths the Lancet field researchers went to or the risks they took to get this information.
Leinad
Fair enough, and good points you raise in addition to being well written etc.
I think the points you mentioned that i didn’t figure was the explanation for excess deaths what it meant. Yes I could well imagine that sewers braking down and causing diseases to infest the population could very well assist in spiking the numbers up.
.
Looking at it that way presents new light I hadn’t considered.
Thanks for explaining it so well we could all understand.
I hope you believe me when I say that I am not trying to cover up for what’s going on etc. I do like getting to the truth of things even if the news is unwelcome and uncomfortable.
Thanks for welcoming me to the hood , zoe . Hope all is well with the little one.
Anyways, no, it wasn’t a “feeling it in the waters thing” as you describe.
It was a statisical inference type thing which I explained why I felt the figures were suspect. There was never any intention by me to argue against the method as I am sure it was well resesrched by superior math wizzes.
to be honest I am highly sceptical of being able to do a ramdom sample in Iraq.
Consider the massive movement in population for a start.
This is as good as one could hope for in that hell hole.
The numbers are silly..Sad as it may be the number of wounded, injured nearly always out number the dead in any bombing (or for that matter fire fight).
So if there is 650,000 dead, where are the millions! of wounded flooding the hospitals.
There is no doubt there is a huge amount of bloodshed and misery but numbers are just not viable.
Oigal: You say there aren’t millions in Iraqi hospitals.
Have you even investigated this?
Have you considered the possibility Iraqi hospitals may be over-tapped already?
Can you think of some reasons why, in the middle of a civil war, people may not be able to access hospital treatment?
Nah. The study must be false, the number just feels too big – right?
PS,
If this went on at my local hossie I’d steer clear too.