Iraq's Official Death Toll Supports Unofficial Tally

This story first appeared at Miller-McCune.

Slow and steady even in tallying the tragic proves its worth.

The government of Iraq has at long last released its own count of violent civilian deaths following the 2003 invasion, and the numbers are close to those from the organization Iraq Body Count.

A story by The Associated Press, which has been hounding the Iraqi government for the numbers, reports that Iraq's Human Rights Ministry tallied a minimum of 85,694 deaths between the start of 2004 to Oct. 31, 2008. The AP, using that source and others, puts the count from the start of the U.S.-led invasion until today at 110,600 dead. Iraq Body Count, in comparison, put its total at between 93,540 and 102,071.

• Counting the dead in Iraq

  For the first time, George W. Bush has announced to the American people the approximate number of Iraqi civilians killed...

• Dep't of Lies, Damn Lies, and Statistics: Iraqi Casualties Edition

  Josh Marshall and the crew at Talking Points Memo are looking into this issue of casualties in Iraq — the...

• Unnamed and Unnoticed: An In-Depth Look at How We Have Treated Iraqi Civilian Deaths

  An in-depth look at how we have treated Iraqi civilian deaths

• Dead Reckoning: Counting Iraq's Civilian Dead

  Meet the men behind the Iraqi casualty numbers.

The Iraqis noted that the real figure may be up to 20 percent higher than their announced figure today. (It also counted 147,195 wounded.) And by comparison, the Web site icasualities.org reports there have been 4,667 coalition fatalities to date, some 4,349 of them American troops.

As we've noted before, counting the dead is fraught with political flak even beyond its macabre, Westmoreland-esque aspects and quantifying human tragedy. No tally, no matter how apolitical, can expect to escape without at least some sniping.

In Iraq in particular, a Lancet study generated lots of attention when it estimated civilian dead through violence at more than 600,000 in the period between the invasion and mid-2006. Iraq Body Count, for example, rejected that count as "improbable" based on both flawed methodology and what might be termed the "smell test."

In The Lancet study, University of London economics professor Michael Spagat pointed out to me, the authors claimed death certificates confirmed much of the carnage they claimed—as high as 70 percent of the violent deaths in 2005. Those numbers just don't gibe—by a factor of between seven and 14—with either the official figures that debuted this week or similar leaked numbers AP reported on earlier.

"Since they claim death-certificate confirmation, they really need it to be the case that there are several hundred thousand death certificates floating around than is actually the case," he wrote in an e-mail. In an as-yet-unpublished paper, he compared the earlier leaked numbers to thoese in The Lancet paper and notes: "It is fairly clear that both the [Lancet] data and the [leaked] data cannot simultaneously be valid.

IBC's own methodology — much more tedious and exacting than the cluster sampling used to produce The Lancet estimate - has shown it holds up. One lesson we can draw is that it is possible to make out important shapes even in the fog of war.