Left I on the News

Didn't I just go over this? Yes, but it looks like I'm not done, thanks to the Associated Press, and their headline which will likely make all the TV news:

AP IMPACT: Secret tally has 87,215 Iraqis dead

But the very first sentence of the article makes clear the headline is false, because that number is the number since 2005 (which is to say it excludes the first three years of the war!) and only applies to those who died by "violent means." And when we get to the third paragraph, we find yet another lie in the headline, since even the authors of the study make very clear their number is "a minimum count of violent deaths." So "Study shows at least 87,215 Iraqis dead from violence since 2005" would be accurate, at least as far as reporting the results of this survey.

AP's article isn't bad as these things go. The authors take pains to note that the results do not include "indirect factors such as damage to infrastructure, health care and stress that caused thousands more to die." They make the magnitude of even their underestimate very clear: "In a nation of 29 million people, the deaths represent 0.38 percent of the population. Proportionally, that would be like the United States losing 1.2 million people to violence in the four-year period." They note that the Iraq Body Count numbers are "even more incomplete" than their study, and they quote one expert as saying that their estimate is likely to be a "gross underestimate." But there are still numerous problems with the article.

The authors actually apologize for the fact that "some slain insurgents could be included in the count," but claim that isn't much of a distortion because "that number was low because few insurgents went to hospitals for treatment out of fear of detection, and many insurgent groups buried their own fighters without getting death certificates." Of course that's precisely the wrong way to look at things. The "insurgents" were, in their vast majority, Iraqis, defending their country from an illegal foreign invasion and occupation; their deaths are the direct result of the invasion and occupation, and not counting them therefore grossly underestimates the number of "Iraqis" who have died.

Then there's the math itself. The authors claim that by adding up the numbers of Iraq Body Count for the war through 2005, the results of this study, and the numbers killed since this study ended in February, they arrive at a total number of Iraqis killed of 110,600 Iraqis. But that's absurd. I can't find IBC's numbers for 2003-2005, but this press release from IBC shows 24,865 civilians killed through March, 2005. Adding in an equal level of deaths through the end of 2005 gives a number 34,189 through the end of 2005. Adding that to the 87,215 gives 121,000, not 110,600.

And finally, we come to the methodology. The authors understand that death certificates are an inadequate method of determining the total number of deaths, and finally, in the final paragraphs of the article (paragraphs unlikely to make it into print when the article is reprinted in most papers, or unlikely to make the headlines on the nightly TV news), the authors get around to mentioning cluster surveys. They refer to the two major ones that have been done, one by the World Health Organization, and the other by researchers at Johns Hopkins University (often referred to as the "Lancet" study). Dealing with the second first, the authors dismiss it as "more controversial." The only thing "controversial" about that study was the result, and the fact that the "establishment" didn't like what it said. Scientifically, it was no more or less "controversial" than the WHO study.

Then there is the WHO study, a subject I dealt with at great length when it was published. The study reported 151,000 Iraqis dead from violence through June, 2006, and also reported a 60 percent increase in nonviolent deaths -- everything from childhood infections to kidney failure -- during the period, an aspect of that study which has been almost universally ignored (except by me). Of course there is no reason whatsoever to exclude those deaths any more than there is a reason to exclude the deaths of insurgents - both were the direct result of the invasion and occupation. AP doesn't bother to extrapolate from that study. I did, in January 2008, and arrived at a figure of 610,000 deaths through that date, 15 months ago.

But "87,000" is the only figure you'll be hearing on the news tonight. Count on it.