Deltoid

ben wrote:

It is a little suspect that this study, like the last, come out right at election time.

Why? There was no evidence that the results in either study were falsified.

Posted by: Robert | October 11, 2006 1:22 PM

It is a little suspect that the rightwingers don't have a counter study to point to because the US military is hiding all their data.

Now why would they do that?

Posted by: melior | October 11, 2006 1:56 PM

Cordesman wrote during the sanctions era that UN studies of deaths in Iraq were so flawed that it raises question to overall level of professional objectivity at the UN. He then went on to cite statistics from the CIA - apparently pulled from the air - showing that by some miracle Saddam had somehow made things better while under the sanctions.

While often informative on the Mid East, Cordesman has no shame in criticising professional studies that don't go his way.

Posted by: Andy Barenberg | October 11, 2006 2:16 PM

He called it a conservative estimate because cluster sampling tends to underestimate where the deaths are radically unevenly distributed. - andy

Posted by: Andy Barenberg | October 11, 2006 3:02 PM

http://littlegreenfootballs.com/weblog/?entry=22920LancetEditor-CertifiedMoonbat&only

Lancet editor is not unbiased. In fact, he has quite an emotional attachment to the whole thing. He comes of as a uber lefty moonbat to me.

Posted by: ben | October 11, 2006 3:49 PM

A friend sent this to me:

My point is, if that many additional people were dying each day for a three-year period, then, it seems to me, that others would have noticed, particularly when you take into account that in large areas of Iraq, particularly the Kurdish area in the north, there is no insurgency, there is little or no sectarian violence,and the population is largely prosperous, semi-independent and self governing. Although not totally peaceful, the areas in southern Iraq controlled by the British and the Poles have experienced far less violence than Baghdad and Al Anbar province, thus strongly suggesting that the death rate in Baghdad and Al Anbar would have to be substantially higher than the average. That is, I submit, sufficient ground to make the report's conclusion inherently suspect.

Does anyone here have a response?

Posted by: Wacki | October 11, 2006 4:28 PM

Wacki, how's this? "The Kurdish area" includes both Mosul, where there has been heavy fighting on and off for three years, and Kirkuk, where there have been well documented allegations of ethnic cleansing. While it's fair to say the Kurdish militias have not challenged American forces directly or joined in the insurgency the way armed men from the other groups have, that doesn't mean the area is entirely peaceful. In the south there are well documented attacks and reprisals between the rival Shiite militias to consider, as well.

The map with the New York Times article shows the distribution of death rates in the Lancet study by province, and does indicate a higher rate in Al Anbar, but non-zero excess death rates in all the provinces sampled.

Posted by: BruceR | October 11, 2006 4:48 PM

Interestingly, the map produced by the study is completely consistent with the relative safety* of the Kurdish areas and southern Iraq.

http://www.nytimes.com/2006/10/11/world/middleeast/11casualties.html

So, rather than discrediting the study, Wacki's wingnut friend is actually confirming its conclusions.

Anyway, according to the CIA World Factbook, the Kurdish people are only 15%-20% of all Iraqis (and not all Kurds live in the "Kurdish area"). So, in terms of the overall statistics, the relatively peaceful* Kurdish area makes very little difference.

https://www.cia.gov/cia/publications/factbook/geos/iz.html

(* Although still more dangerous than almost anywhere in America)

Posted by: theo | October 11, 2006 5:36 PM

Wackj: Considering the confidence interval was 8,000-194,000 I find it very difficult to understand how this guy came to that conclusion. Does anyone have this quote in context? This guy has certainly made some statements that just sound fishy to me.

1. "Survivor bias" - where entire households died there was no-one left to report.

2. The checking of death certificates made it difficult for people to fabricate additional deaths but it was a lot easier for them to lie and understate deaths - say because they were afraid of reprisals if a household member had been an insurgent.

They also threw out a cluster from Fallujah showing an extremely high death toll there because they were afraid it would bias the figure upwards. That meant Anbar, the most violent part of Iraq according to most reports, was underrepresented.

Posted by: Ian Gould | October 11, 2006 7:03 PM

The question implied by the headline of this post, and still unanswered is:

How many innocent Iraqis have to die before Americans realize this war is a bad idea?

Posted by: Spike | October 11, 2006 9:15 PM

Hi all Like I have said B4, when sandniggers wipe each other out it is a doule whammy. Islam is a beleif of death, long live white anglo saxon values. Regards Adolf Powell

Posted by: Adolf Powell | October 12, 2006 6:11 AM

Malcolm Ritter in the Associated Press did a hatchet job on the study...In the second word of his story he tags it as "controversial",

Well, duh. People are indeed arguing over it, so by definition it's controversial. It's rather arrogant for you to consider it gospel truth simply because it confirms your own viewpoint.

and in the second sentence he has an expert dismiss it as "politics". And how can the timing of the release possible cast doubt on the accuracy of the study?

Because they're releasing it just three weeks before mid-term elections in America, just like they released their last controversial study right before the presidential election. Likewise, consider the fact that the study includes deaths only tangentially related to the invasion, and that it's overall figure doesn't distinguish between those killed by the Coalition and those killed by insurgents. By this we're to assume that the Coalition is ultimately responsible for killings they don't commit, and in fact try nobly to prevent every day. This sort of assumption is both highly debateable and highly political, and makes the report itself a political document.

And, as evidenced here and elsewhere, most liberals share a visceral hatred of George W Bush and the Iraq war (even though it originally had bipartisan support based upon a consensus opinion about the threat Iraq). Given this blind hatred I don't put it past anyone to lie or exaggerate in order to hurt the war effort. Afterall, what's more important? Being totally honest, or stopping the deaths of civilians and defeating old Georgie.

So who's the expert?

Obviously not you, a point made hilariously obvious by the fact that you don't know how to read a paper.

http://timblair.net/ee/index.php/weblog/paystocheck/

But I will give you one point which you make particularly about Americans, but is universally true. People are willing to accept large numbers of casualties if they think that their cause is just, and more importantly, victorious. Americans perhaps understand this better than Australians because we actually fought for our independence rather than begging it, but I still believe that victory trumps other considerations for most people in the world. Conversely, failure drains the moral worth of even the noblest cause, which is why most people are now against the war. But you should be careful, for if we do establish a stable democracy in Iraq most will remember only this: we won fighting for a just cause, and you opposed it.

Posted by: chairman me | October 12, 2006 11:33 AM

"A controversial new study contends nearly 655,000 Iraqis have died because of the war, suggesting a far higher death toll than other estimates."

Yeah, it's "controversial" because the method they're using to come up with these absurd numbers has already been demonstrated to produce totally unreliable results, as per the FAO Baghdad infant mortality study conducted a few years back (also published in The Lancet) in which it was later discovered that MOST of the people who were being surveyed simply lied in regard to questions about family members who had died.

The fact is is that these survey methods are a total joke, the numbers they prodiuce are a total joke and, what's more, the people at The Lancet know they're a total joke.

Posted by: Dave Surls | October 12, 2006 1:03 PM

Sorry, JohnP, you're estimation is flawed in at least two ways. First, it assumes that the violence nationwide is occuring at the same rate as Baghdad, when Baghdad is the currently the most violent city in Iraq. Second, it assumes that the level of violence has been the same since day one--funny, given that the left loves to always describe the violence as "spiraling out of control" like a broken jetski or something.

Inadvertantly,though, you demonstrate very clearly how one's own assumptions can color statistics, which is often described as more art than science.

Posted by: chairman me | October 12, 2006 1:59 PM

Baghdad is nowhere near the most violent city in Iraq.

I believe that in terms of civilian deaths it is, but I'm not claiming to be an expert, so please explain the claim if it's true.

Update: In comments, ragout informs me that the story was referenced on the front page. The actual story, however, was not on the front page, but buried on page 12.

That is the lamest, most weasily defense I've seen in some time. So because they only included a huge picture of an Iraqi woman with a coffin plus a prominent referral to page A12, you consider it buried. You know sometimes it's better just to admit you did something boneheaded instead of defending it in such a way that makes you look ridiculous.

Posted by: chairman me | October 12, 2006 3:05 PM

I believe that in terms of civilian deaths it is, but I'm not claiming to be an expert, so please explain the claim if it's true.

Some quick Googling turns up references to both Falluja and Ramadi as Iraq's most violent cities. On what are you basing your belief that Baghdad is the most violent?

Posted by: Davis | October 12, 2006 4:29 PM

Dave, what peer-reviewed papers make that claim? As has been pointed out, death certificates were produced in the majority of cases.

Posted by: Barry | October 12, 2006 5:17 PM

Tim,

And the innumerate Lancet critics will be [admitting they're wrong] when?

Innumerate? Well, actually, I think I can give you a rough estimate of the actual number of Lancet critics. I've spoken with 10 people about the Lancet study, and 7 of them think it's utter bullshit. Extrapolating to the population of Earth, that means about 4 billion of us think the study is wrong. And you know what, The Post has totally ignored this important fact!

As for admitting a mistake in questioning the Lancet study, that really depends on whether or not someone can defend it coherently, and I really doubt that someone is a guy who doesn't know how to even read a newspaper.

Davis,

I'm simply basing it on news stories I read coming out of Iraq that suggest most of the civilian deaths are occurring in Baghdad. I don't know if that's a fact or not, or if Ramadi has more violence and less civilian casualties. Incidentally, most of the fighting in Anbar province now concerns Coalition forces and Sunni partisans fighting al Qaeda and one another, while the violence in Baghdad seems to specifically target civilians. My point, though, was that interpolating an accurate body count from a particularly violent city will inevitably make the count much higher than reality.

Posted by: chairman me | October 12, 2006 6:27 PM

"My bad, Dave. I accorded you some respect."

That's not bad. You were wise to do so.

Posted by: Dave Surls | October 12, 2006 6:59 PM

Gee, I'm glad Doug didn't sully that pristine comment with, y'know, content.

Posted by: RobW | October 13, 2006 4:45 AM