Science-Based Evidence

I have no idea whether the Lancet study is right or wrong.  And neither does anyone else, including the authors. 

~Megan McArdle, aka Jane Galt

That's my favorite quote on this year's attempt to damage Republican chances in an election with an Iraq death count study, brought to you by the same folks who tried do that last time. Not that columnist Eugene Robinson seems interested or aware of that fact:

...quite a lot of science went into the Johns Hopkins study. Even if you assume that the number of Iraqi civilians killed since the war began is at the very low end of the study's range, that's still a quantum leap from earlier estimates. We now have reputable evidence -- not proof, I'll allow, but science-based evidence from respected scholars, published in one of the world's most prestigious medical journals -- that the humanitarian tragedy in Iraq is much, much worse than anyone had suspected.

Yesterday I briefly noted the abysmal weakness of the first study by Gilbert Burnham and Les Roberts, which claimed a civilian death toll of about 100,000 at an earlier stage in the war. One of the many problems with the new study, claiming 655,000 "extra" deaths as a result of the war, is indeed that it's a quantam leap from earlier estimates but I think Robinson says that's not a bug, it's a feature.

Clayton Cramer observes:

The more I think about the Lancet article, the more obviously bogus the results are. The claim is 654,965 excess deaths caused by the war from March 2003 through July 2006. That's 40 months, or 1200 days, so an average of 546 deaths per day....

Some back-of-the-envelope work shows this would require some days with 10,000 deaths or a lot more with 1,000 or so. He continues:

So, where are the news accounts of tens of days with 10,000 or more deaths? Where are the news accounts of hundreds of days with 1000 deaths or more? This article claims that there are perhaps 100 Iraqis a day now being killed in sectarian violence--and this is described as escalating violence.  This horrifying article talks about 65 bodies found around Baghdad--with the claim that the day was "notable in its number."

Either the news media have been ignoring hundreds of days with 1000 or more deaths--or tens of days with 10,000 or more deaths--or the Lancet article is utterly wrong.

Meanwhile Omar at Iraq the Model is not impressed with the authors' concern for Iraqi lives:

We believe in what we’re struggling for and we are proud of our sacrifices....

To me their motives are clear, all they want is to prove that our struggle for freedom was the wrong thing to do. And they shamelessly use lies to do this…when they did not find the death they wanted to see on the ground, they faked it on paper! They disgust me…

This fake research is an insult to every man, woman and child who lost their lives. Behind every drop of blood is a noble story of sacrifice for a just cause that is struggling for living safe in freedom and prosperity.

UPDATE: Mark Finkelstein at Newsbusters has an outstanding post that documents objections to the "science-based evidence" of this study, the Democratic, antiwar political sympathies of at least one of its authors, and the rhetoric, worthy of Michael Moore, of the Lancet's editor. Here is Richard Horton,  editor of what Eugene Robinson calls "one of the world's most prestigious medical journals":

Lancet editor Richard Horton, speaking at an anti-war rally last month, said that “[the British] government… prefers to support the killing of children instead of the building of hospitals and schools…As this axis of Anglo-American imperialism extends its influence through war and conquest, gathering power and wealth as it goes, so millions of people are left to die in poverty and disease.” Hat tip Little Green Footballs.

Though I hear at one time The Lancet really was prestigious.

Comments

[paul zummo]

Eugene Robinson is by leaps and bounds the dumbest and most intellectually dishonest major newspaper coulmunist I have ever read, and that includes the likes of Dowd and Krugman.

[Engineer]

1 dead for every 4 randomly selected home. That's bad no matter how you look at it. (They interviewed 1,840 random people and found over 547 dead - 92% of those showed the death certificate)

If you think about it we've dropped over 240,000 cluster bombs. We'd be fools to think they didn't kill anyone. Add in gunfire and car bombs and 600,000 dead doesn't seem that big.

[Engineer]

1 dead for every 4 randomly selected home. That's bad no matter how you look at it. (They interviewed 1,840 random people and found over 547 dead - 92% of those showed the death certificate)

If you think about it we've dropped over 240,000 cluster bombs. We'd be fools to think they didn't kill anyone. Add in gunfire and car bombs and 600,000 dead doesn't seem that big.

[ErnestD]

Paul Zummo, take a look at the figures again.

They didn't interview "1840 random people;" they sampled 1849 households that were grouped into 47 clusters. There were a total of 12,801 people in these clusters.

The survey teams "randomly" chose these 47 locations, leaving out two of the least violent Governorates in Iraq, then went house to house until they had interviewed 40 households in the same little cluster area. As even the authors admit, "the potential exists for interviewers to be drawn to especially affected houses through conscious or unconscious processes." Say, for example, because they are opposed to the US occupation and know that the people who are paying them, the survey supervisors, are outspoken anti-war activists who already believe the US should leave Iraq and wish to help change the public perception of the war.

(The authors also apparently undersampled rural regions, leading to inflated violent death numbers and an artificially low non-violent death baseline.)

Second, the researchers recorded a death certificate as "present" in only 80% of the total deaths, not 92% as stated misleadingly in the article. Only 545 certificates were requested, and 501 of these were "present"...but researchers did not ask for certificates in the remaining cases (total deaths=629). As for whether each certificate was examined rigorously and tested for authenticity...the article isn't clear on this and I wouldn't hold my breath.

This lack of rigor is, regrettably, common in epidemiology and also largely unavoidable. It happens when a Western researcher with little or no local linguistic or cultural fluency uses a local research team to investigate a controversial topic in an underdeveloped country. The data are unverifiable and amount to hearsay. The subjects tell the survey team what they think they want to hear, and the survey team consciously or unconsciously directs the survey and redacts/compiles the results in such a manner as to obtain what they think the Western supervisors want to hear.

[ErnestD]

Sorry, Paul, I meant Engineer.