On civilian deaths in Iraq

On civilian deaths in Iraq
by Wm. Robert Johnston
23 October 2006

With U.S. midterm elections just 27 days away, the British medical journal Lancet has released a study claiming that 655,000 Iraqi civilians have died due to the U.S.-led invasion and occupation of Iraq. This study has been warmly received by opponents of the invasion of Iraq, despite significant reasons to question the validity of the estimate. Perhaps more amazing than the claim itself is the credibility granted to a study flawed in so many ways.

The researchers, Burnham, Lafta, Doocy, and Roberts, decided they couldn't trust actual counts of those killed. Rather, they enlisted workers to go door-to-door to 1,849 households in 47 Iraqi neighborhoods. From this sample of 0.05% of Iraq's population, they extrapolated the 547 reported post-invasion deaths to a total for the country. Further, they subtracted from their derived death rate their assumed value for the pre-invasion death rate.

Burnham et al.'s stated reason for such an approach is that traditional reports of Iraqi deaths are incomplete. In such a case, valid statistical sampling can provide an alternate measurement. While other scientists have been cited as affirming their methods, there are some qualifiers for this study to be scientifically valid: the sample must be representative of the whole population, the underlying assumptions must be valid, and the estimate must be reproducible.

The fact that the study's results are not reproducible, that they are so far removed from other estimates and measurements of Iraqi deaths, should give the reader pause. The Iraqi Ministry of Health reports about 40,000 violent deaths, based on death certificates. The leftist group Iraqi Body Count reports 44,380 to 49,297 "civilian deaths resulting from the US-led military intervention in Iraq" through 22 October 2006 based on media reports. Data released by the U.S. Defense Department implies about 54,000 Iraqi casualties (both killed and wounded, both civilian and Iraqi security forces) through 11 August 2006.

The Lancet study claims these estimates are invalid due to underreporting. But even surveys using the same sampling methodology but with larger samples, give estimates more consistent with the smaller estimates. A United Nations Development Programme survey of 21,668 Iraqi households estimated between 18,000 and 29,000 violent deaths through mid-2004. Another "study" cited by Burnham et al. is one attributed to Iraqiyun claiming 128,000 deaths. It is not clear whether or not this "study" includes deaths from natural causes, or even if it is much more than an Internet rumor.

The representativeness of the Burnham et al. sample is also dubious. Their survey teams avoided back alley areas for safety reasons. By avoiding areas likely shielded from the car bombings typical on main streets, however, they could have introduced bias. The surveys were limited to urban areas, with some provinces represented by a single "cluster" of 40 households. Some researchers have questioned how reliable these results are, with typically 40 households surveyed by a team in a single day. Others have questioned how open and honest Iraqi civilians would be with total strangers on this subject.

Burnham et al. obtained a figure of 655,000 excess deaths since the U.S.-led invasion, or a range of 393,000 to 943,000, with 601,000 of these deaths due to violent causes. But this assumes a pre-war mortality rate. Again, rather than use measured values, they use a value of 5.5 per thousand per year derived from their survey. This is certainly biased towards lower values for a variety of reasons, and comparison to published mortality rates affirms this: the CIA reported a rate of 5.84 in 2003, and the United Nations Population Division gave a rate of 9.7. Simply using the UN mortality rate would reduce the Lancet result to 300,000 excess deaths.

But a more insidious flawed assumption involves ignoring unreported deaths under Saddam Hussein. Official reports of death rates in pre-invasion Iraq don't include deaths due to government repression. Human Rights Watch estimates that 290,000 people were "disappeared" under Hussein. In 1988 alone Hussein's government killed an estimated 100,000 Kurds. Recently, French human rights activists published a book claiming Saddam Hussein's government killed as many as 2,000,000 Iraqis during its reign. Since the 2003 invasion, hundreds of mass graves from the Hussein period have been unearthed, some containing dozens, some thousands of bodies.

One could also question the implied attribution of blame to the U.S. and its coalition. Most of the violent deaths occurring in Iraq are at the hands of the Islamic terrorists who seek to rob the Iraqi people of the democracy they so clearly desire. Even the Iraqi Body Count estimates only attribute 37% of civilian deaths through March 2005 to coalition military action, with the overwhelming majority of these occurring during the invasion itself.

Could the Burnham team been willing to publish such unsupported results for political reasons? After all, the study was published just before U.S. midterm elections. We are assured that this does not undermine their "scientific" results, just as with the previous release by the same researchers: a Lancet study released 4 days before the November 2004 U.S. presidential election, claiming that 98,000 Iraqis had died due to the U.S. invasion (although they admitted the true figure could be anywhere from 8,000 to 194,000, given their small sample size).

The irony is that some seem to feel a need to exaggerate reality, as if the real death toll was not tragic enough. The reality, that perhaps 40,000 Iraqis have died due to violence during the occupation period, nearly all at the hands of terrorists and criminals, is nothing to scoff at. This has not deterred the Iraqi people from participating in elections, from seeking the fruits of this struggle. At the same time the U.S. news media has a fixation on the toll among U.S. troops: 2,791 killed in Operation Iraqi Freedom through 23 October 2006. The sacrifice of these soldiers is demeaned by this fixation: hundreds of thousands of U.S. troops, some making the ultimate sacrifice, are laboring professionally to save the Iraqi people from oppression and to secure Americans and others from terrorism.

But this is not tragic enough for some. We can make one projection based on these studies: just before the 2008 U.S. presidential election, the Lancet will release a study claiming that the U.S. has killed millions of Iraqi civilians. After all, there is a noble political agenda here, criticizing the United States; for that, we don't need no stinkin' facts.