Slate's war on epidemiology

You would have hoped the editors of Slate would have taken into account the way Fred Kaplan's innumerate criticism of Lancet 1 was shredded, but they've gone and published an attack on the study by Christopher Hitchens, who knows less about random sampling than Kaplan. I already caught Hitchens lying about the first Lancet study, so you can probably guess what it's going to be like. Let's go:

The word lancet means either an old-fashioned surgical knife used to open a vein for the once-popular cure-all remedy of "bleeding" or "bloodletting," or (in architecture, especially Gothic) a rather narrow window. Both metaphors seem apt for the British medical journal of the same name, which appears to be seeking a reputation for conjuring bloodbaths and then reviewing them through a slitlike aperture.

We used to do this in high-school debates. Open with a dictionary definition. Seems kind of lame now. (Yes, I was on the debate team. What did you expect?)

In December 1995, the Lancet published another equally disturbing document, this time a letter to the editor from Sarah Zaidi and Mary C. Smith Fawzi. They relayed the findings of a study they conducted for the U.N. Food and Agriculture Organization that estimated that 567,000 Iraqi children had died "as a consequence" of sanctions imposed on Iraq by the United Nations in August 1990. Note, again, the slightly subjective definition of cause of death.

That was a letter to the editor, not a refereed article. And in October 1997, the Lancet published a letter from Sarah Zaidi who withdrew the claim after a follow-up study found a much lower numbers of deaths -- the 1995 survey found an 20% child mortality rate, when repeated in 1996 the rate was 3.8%.

But it does seem, according to the Lancet, that hundreds of thousands of Iraqis were doomed to die, one way or another, in peace or in war, unless Saddam was left unmolested. Since the signature features of Iraq under the Saddam regime were the killing field abroad and the mass grave at home, this seems to leave few good options.

Actually that should be "according to one letter to the Lancet which was subsequently retracted". And the oil-for-food program ended the harmful effect of sanctions on Iraqi health, so no, they weren't doomed.

There have been several challenges to the epidemiology of the Lancet/Johns Hopkins team concerning their definition of a population sample.

In the original the link went to the IBC's attack which had nothing relevant to say about the epidemiology.

But I see no reason in principle why anyone who endorsed the liberation of Iraq, and who opposes the death squads of the Baathist/jihadist "insurgency," should want or need to argue that the casualty figures are any lower.

Hundreds and hundreds of warbloggers disagree.

If the cause of all this death is "the war," does that mean that the coalition has killed nearly 700,000 Iraqis? Of course it means nothing of the sort. Indeed, if you look more closely, you will see that less than one-third of the surplus deaths are attributed, even by this study, to "Allied" military action. Grant if you wish that this figure is likely to be more exact, since at least the coalition fights in uniform and issues regular statistics.

"Regular statistics" on how many people they've killed? I don't think so. And the coalition is responsible for the deaths that result from their war of choice.

That leaves, according to the Lancet, a pile of corpses nearly half a million high. Here, the cause of death becomes suddenly less precisely identifiable. We are told that 24 percent of the violent deaths were caused by "other" actors, and 45 percent of them by "unknown" ones. If there is any method of distinguishing between the "other" and the "unknown," we are not told of it.

Except, that is, in the study. "Other" deaths were attributed to non-coalition actors. "Unknown" means that the respondent didn't know whether or not the coalition was responsible.

Make the assumption that some percentage of those killed by the coalition are the sort of people who have been blowing up mosques, beheading captives on video, detonating rush-hour car bombs, destroying pipelines, murdering aid workers, bombing the headquarters of the United Nations, and inciting ethnic and sectarian warfare. Make the allowance for the number of bystanders and innocents who lost their lives in the combat against these fanatics (one or two, alas, in the single case of the precision bombing of Abu Musab al-Zarqawi, just to take one instance). But who is to say how many people were saved from being murdered by the fact that the murderers were killed first?

Not to mention that some of those innocent children blown the hell up might have grown up into murderers. This is just pathetic weaseling. The Lancet estimate is the loss of lives on net. How many more violent deaths there were than before the war. A war which the coalition chose to start.

But the "tit for tat" confessional killings were and are a deliberate tactic of the insurgency and now threaten to spread into mass reprisals on both sides, while all the effort of the coalition is devoted to negotiating a compromise between the country's factions. It is simple moral idiocy to fail to distinguish between these phenomena.

It is, of course, Hitchens who display moral idiocy here. First, the study did try to distinguish between coalition-caused deaths and those caused by other people. Second, all these people are dead because of a war of choice, a war that Hitchens loudly supported. Hitchens displays moral idiocy in his attempts to shrug off responsibility. The warbloggers frantically scratching around for reasons to discount the Lancet study display a greater moral sense than Hitchens.

[Saddam's] successors and allies did not allow one day of peace after the invasion before launching a hair-raising campaign of murder and sabotage and consciously inciting a civil war.

What was Hitchens expecting them to do? Play nice?

Lancet figures are almost certainly inflated, not least because they were taken from selective war-torn provinces.

And also from non war-torn provinces. Provinces were sampled proportional to their population, so this would not inflate the figures.

But there is no reason why they may not come to reflect reality more closely. It is a reminder of the nature of the enemy we face, and not only in Iraq, and a very clear picture of the sort of people who would have a free hand in Iraq if the coalition were to depart.

So the more deaths there are, the more justified the war is. Moral idiocy indeed.

Steven Poole has more on Hitchen's moral idiocy.

And this is what Slate has at the bottom of the page.

Fred Kaplan discredited an earlier report.

Did not. (Update: Slate changed it to say "disputes" instead. It still says that in another article Kaplan "discredits" David Kay's assertion. Slate therefore concedes that Kaplan did not discredit Lancet1.)

Update Robert Farley:

We're obviously beyond the point where one could say with any degree of originality that Christopher Hitchens is a morally and intellectually bankrupt sociopath. He is the true heir to the Stalinist left that he relentlessly rails about; there is no limit to the death and destruction that he's willing to tolerate in service of his revolution. What's more important now is to note that those who willingly associate themselves with people like Hitchens and Bill Kristol should be viewed in the same light. To paraphrase Yglesias, even if we were to find something of value in the Euston Manifesto or the work of PNAC (and this is a tremendous "if"), associating with the people who press these intellectual projects is, in itself, evidence of a lack of seriousness about foreign policy.