The October online posting by WikiLeaks of nearly a thousand classified Pentagon documents (the “Iraq War Logs”) shed new light on the vexed issue of Iraqi deaths during and after the 2003 invasion. According to articles in the New York Times, Los Angeles Times, and elsewhere, the Bush administration, despite its claims to the contrary, did in fact keep a running count of Iraqi fatalities, entering 66,081 civilian deaths into the Iraq War Logs between January 2004 and December 2009 (out of a total of 109,032 recorded violent Iraqi deaths).
Yet the war logs yield no clear evidence that the U.S. government made any systematic effort to record all Iraqi deaths. Reports of Iraqi fatalities seem to have entered almost inadvertently into the Pentagon’s files as part of the daily bureaucratic grind of the occupation. General Tommy Franks’s now infamous declaration, “We don’t do body counts,” remains true, by every indication, as far as U.S. war policy is concerned.
This means that those 66,081 civilian deaths entered into the war logs likely represent a great undercounting of actual deaths. There are several reasons why. First of all, a large number of killings would have occurred in the nine months of the invasion prior to January 2004, when the war logs paper trail begins. Many deaths, moreover, would have occurred in sectarian violence in remote or uncontrolled parts of the country beyond the view of U.S. forces. From what we now know of the conduct of the war, some deaths labeled by the military as combatant deaths were assuredly civilian deaths, including individuals killed in what amounted to “free fire” zones (such as the city of Fallujah in November 2004). And still more deaths may well remain hidden in secret documents we have not yet seen.
The total number of civilian deaths, many journalists now report, should be placed in the range of 122,000—an estimate based on the WikiLeaks revelations combined with earlier tallies by the Iraq Body Count (IBC), an independent British-based group that began tracking news stories of Iraqi civilian deaths in March 2003. Yet the IBC, as anyone scrutinizing the group’s stated goals and methods will discover, makes no attempt to track the total number of civilian deaths from the war. In all likelihood, the IBC’s updated figure of more than 122,000 reliably documented violent Iraqi civilian deaths is hundreds of thousands lower than the number of actual deaths from all causes of the war.
The question of Iraqi fatalities is, of course, a highly contentious one, and those in this country who seek a sober estimate risk being accused of politicizing the war or dishonoring U.S. service members. But political as well as psychological temptations to suppress or distort the facts also burden those who supported the rush to war and who thus share moral responsibility for the human catastrophe that followed. Whatever one’s political commitments, facing the question of Iraqi civilian deaths as honestly and objectively as possible is both an intellectual and a moral imperative.
Several rigorous attempts to count the human costs of the war in Iraqi lives have been made. The three most important are the ongoing IBC tabulation; a survey conducted jointly by researchers from Johns Hopkins University, MIT, and Al Mustansiriya University in Baghdad, published in the Lancet in 2006; and the Iraq Family Health Survey (IFHS), administered by Iraq’s Ministry of Health and published in the New England Journal of Medicine in 2008. Each of these studies has serious methodological weaknesses, and their respective authors have engaged in acrimonious disputes with each other over assumptions, approaches, and conclusions. Nevertheless, all agree that the number of fatalities from the war is vastly higher than 122,000. The only question is how much higher.
The IBC’s figures derive not from its own fieldwork but rather from meticulous collation and analysis of publicly available reports from NGOs, hospitals, morgues, and—primarily—Western news articles. The group gathers reports detailing violent civilian deaths from media outlets that have achieved “a respected international status” and that maintain rigorous standards of fact-checking. It then sifts these stories to create a single database of individuals killed. In order to be entered into the IBC’s tally, the same fatality must appear in a minimum of two independent sources (with some “provisional” exceptions, according to IBC’s Web site). As of November 2010, the IBC reported between 98,585 and 107,595 civilian deaths from war-related violence, not including the latest WikiLeaks disclosures. The range reflects conflicting reports of, for example, whether those killed were civilians or combatants. The IBC makes clear, however, that its figures are not a count of all Iraqi deaths. Its database provides a record only of credibly documented or known civilian deaths from war-related violence. “Our best estimate,” IBC co-founder John Sloboda told the BBC in 2006, “is that we’ve got about half the deaths that are out there.”
The Bush and Blair administrations dismissed the IBC’s work, arguing that media reporting of civilian deaths is simply unreliable. Other critics of the IBC’s methods have questioned the reliability of accounts posted by reporters who for reasons of personal security never strayed far from Baghdad, and then were almost always “embedded” with the U.S. military. Furthermore, the IBC analyzes only reports by journalists in the English-speaking press. Articles documenting civilian deaths by journalists working in Arabic, French, German, or other languages are not considered by the IBC unless these stories are first translated into English by others and/or re-published in English-language media. The organization, some have suggested, might therefore more accurately be called the “Iraq English Media Body Count.”
There exist other reasons to suspect that the IBC count might be low. Through much of the war, journalists tended to focus their reporting on spectacular, politically significant, or large-scale incidents of violence. Isolated killings of ordinary people in the confusion of U.S. military raids and roadside checkpoints—and more importantly, in the waves of sectarian violence unleashed by the invasion—were rarely deemed newsworthy events (let alone reported by more than one independent publication, as the IBC requires). Yet these forms of “mundane” violence might well have claimed many thousands of lives. If there were on average three killings a day from sectarian violence in each of Iraq’s seventy-five major cities, the authors of the 2006 Lancet study pointed out, this would add up to more than two hundred and seventy thousand deaths in a little over three years of bloodletting.
Is it plausible that violence of this magnitude occurred throughout Iraq over an extended period? An Associated Press study of files from the Baghdad, Karbala, Kirkuk, and Tikrit morgues for the period May 1, 2003 to April 30, 2004 found that they had collectively recorded 5,558 violent deaths—an average of 3.8 violent deaths per day per morgue. (And remember, April 2004, the month the AP study ended, was the month the Shiite sectarian uprising led by radical cleric Muqtada al-Sadr began.) Granted, there would presumably have been less violence in small cities than in the four major centers in the AP report. But then, not all persons killed in fighting were delivered to Iraqi health officials at the country’s morgues either, since Muslim tradition dictates that bodies be buried as soon as possible, with every effort being made to do so within twenty-four hours of death. Nor were all deaths in the war caused by violence or fighting. “Report based” methods of counting fatalities such as the IBC’s tell us nothing about unreported killings and also nothing about the overall picture of what epidemiologists refer to as “excess mortality”—that is, premature deaths, ones that happen before the average life expectancy for any particular demographic category.
When a country’s infrastructure is destroyed, depriving people of clean water and electricity, and making travel so dangerous that they are unable to visit clinics and hospitals, death rates climb in ways soldiers or reporters might never detect. But public-health experts point out that these, too, are casualties of war. The 2006 Lancet study (building on an earlier 2004 survey by the same authors) was the first scientific attempt to determine how many Iraqis had died from all causes during the war. It was also the first analysis of Iraqi war deaths to be published in a peer-reviewed academic journal. (The Lancet is one of the world’s oldest and most prestigious medical publications.) The study followed well-established research methods that have been used without controversy in other conflict and disaster zones where data collection is extremely difficult, such as Darfur, the Congo, and the countries affected by the 2004 tsunami.
The approach—known as “random cluster sample surveying”—involves identifying “clusters” of households that are representative of a country, weighted for population density in different regions. Interviewers travel to these households and ask detailed questions about deaths in their families. Responses are then used to extrapolate mortality figures for the country as a whole. The more households included in a survey, the more confident one can be about the results. The Lancet study—which included interviews of 1,849 households containing 12,801 persons, a respectable sample size—found that in the first forty months of fighting, between 392,979 and 942,636 Iraqis had died from all causes of the war. The wide range, or “confidence interval,” is what one would expect given very different rates of violence in different regions of the country. The number of “excess” deaths, the authors reported, was most likely 654,965, with more than 90 percent of these due to violence.
Not surprisingly, the finding that well over half a million Iraqis had died violent deaths in the first three years of the war was greeted with incredulity, anger, and scorn by many cheerleaders of the invasion. Perhaps the most invidious line of attack was the one promoted by, among others, conservative pundit and Fox News commentator Michelle Malkin, who featured a cartoon on her Web site in which a man in a lab coat labeled “Lancet Survey” interviews an Iraqi couple in their doorway. “Scientifically speaking,” he asks the pair, “how many Iraqis would you say have died since the invasion by the blood-sucking axis of Anglo-American imperialist cluster-bombing baby-killers?”
Such crude efforts to discredit the Lancet study by depicting its authors as “unpatriotic” in the “war on terror” were clearly nothing other than an ad hominem attack. With nearly two-thirds of the total number of violent deaths in the Lancet survey attributed not to coalition forces but to fellow Iraqis, it would make as much sense to label the study “anti-Iraqi” as “anti-American.” And while it is certainly possible that some people lied about deaths in their families out of resentment toward the United States, it is equally possible that people underreported violent deaths by U.S. forces out of fear of retribution in cases where family members had been involved in violent activities against coalition troops. Tellingly, none of the most vituperative critics of the Lancet article urged the U.S. government or the many well-funded conservative think tanks to replicate and refine the survey in order to arrive at a more accurate picture of the human costs of the war. Having a serious public conversation about Iraqi deaths was clearly something many armchair statisticians wanted to avoid at all costs.
Ironically, after the Lancet article was published, the IBC—whose figures had earlier been attacked in similar terms—suddenly became the preferred authority among conservative commentators taking on the question of Iraqi deaths. In 2005, the National Review had dismissed the IBC as a “hard left” group “blindly throwing darts at a dartboard.” By 2007, David Frum in the National Review was using the figures of the IBC—now described simply as an “antiwar Web-based charity that monitors news sources”— to discredit the Lancet’s findings. By 2008, National Review columnist Michael Rubin was directing his readers’ attention to a “damning” and “great” article on Iraqi civilian deaths published in the National Journal. The National Journal article critiqued the Lancet study in contrast to what it described as the IBC’s “cautiously compiled database.” The IBC in fact continues to be widely cited by reporters across the political spectrum as the most realistic source on Iraqi casualties. The Lancet survey, with its higher death totals, is presented as unreliable in comparison—or, more often, simply ignored. Since 2004, the New York Times has cited the team’s work in its reporting and op-ed pieces thirteen times, once on the front page. In the same time period, the Iraq Body Count has been mentioned thirty-four times, eight times on the paper’s front page.
Yet Malkin’s insinuation that the teams who administered the Lancet survey were effete leftists (if not Islamist ideologues) bent on tarnishing America’s image is itself the sure sign of an ideologically driven true believer. It is also a shameful response to Iraqis who put their lives in peril to provide information to the world about an unfolding human disaster in their country. There is no credible evidence whatsoever of ideological bias in the Lancet study. Its authors were, if anything, excessively conservative and cautious in their approach, as David Marker points out in an insightful review in the Public Opinion Quarterly. For example, they assumed that several regions of the country where they were unable to safely travel had zero violence, and they excluded the devastated city of Fallujah entirely (as an abnormally violent outlier).
There are, however, several possible sources of significant bias in the Lancet study, which suggest that its figures should be treated with extreme caution. One involves a sampling error that the authors were unable to control for. The cluster selection the Lancet team used was based on two-year-old information about Iraq’s population distribution—the most up-to-date information then obtainable. But the war caused a massive displacement of people from their homes. According to the U.S. State Department, there are still more than 2 million Iraqi refugees living overseas and another 2.8 million displaced persons within the country. If there were 2 million fewer people in the country than the Lancet study assumed, and if people fled more violent provinces at very high rates, this would result, according to Marker, in an overestimation of mortality in the Lancet study by perhaps as much as 10 percent.
Pointing out possible sources of error is not the same, though, as disproving (let alone improving) a social-scientific study. To do this, one would need to come up with a better research design and then test it through the hard—and in the case of Iraq, often deadly—task of new data collection. This is exactly what the Iraq Family Health Survey (published in 2008 but covering the same time period as the Lancet study) attempted to do—in the process suffering the loss of a team member shot and killed while conducting interviews.
The IFHS study was funded by the United Nations and administered by the Iraqi Ministry of Health, with technical assistance from the World Health Organization (WHO). It followed the same cluster sample method used by the Lancet team, but increased the sample size by about fivefold. Published in the New England Journal of Medicine—like the Lancet a peer-reviewed journal—the study set a range of between 104,000 and 223,000 violent deaths from the invasion through June 2006, estimating an exact number of 151,000. However, it reported an additional 250,000 deaths from nonviolent causes, making for a death total of approximately 400,000.
The Lancet authors greeted the IFHS figure as a confirmation of their own, since 400,000 deaths fell within their confidence margins. At the same time, they criticized the -IFHS’s work on several grounds. Both the IBC and the Lancet survey showed a dramatic increase in violent deaths of Iraqis in 2005 and 2006. The IFHS study, however, showed an improbable steady rate of deaths for all years. The Lancet authors argued that IFHS’s interviewers had failed to gain accurate responses to questions dealing with political or sectarian violence as a cause of death.
One serious problem with the IFHS survey was that its interviewers introduced themselves to households as representatives of the Ministry of Health (MoH). At the time, the MoH was under the control of forces fiercely loyal to Muqtada al-Sadr—a man responsible for orchestrating much of the violence the IFHS teams were seeking to measure. MoH hospitals were “de facto outposts for Shia militias,” the London Times reported in February 2007. Reports told of Sunnis being refused treatment at MoH facilities or even shot in their beds. Deputy health minister Hakim al-Zamili was arrested in early 2007—while the IFHS interviews were still being conducted—on charges of, among other things, using MoH ambulances and hospitals to carry out killings, and plotting the kidnapping and murder of his Sunni colleague, Ammar al-Saffar. Another Sunni health official, Ali al-Mahdawi, who ran the Diyala Health Directorate, had “disappeared” along with several of his staff after they traveled to MoH headquarters in Baghdad for an interview several months earlier.
In light of these facts, the IFHS conclusions must be read with great caution. Did the teams systematically underreport Sunni fatalities? Conversely, did they inflate Shiite deaths? Even if the surveyors performed their work with integrity and rigor under the careful watch of the WHO officials who helped design the study, did the mere fact that they were from the Ministry of Health influence the answers they received? If a family’s son had been murdered by Shiite militias, for example, and a group of interviewers from the MoH appeared at their doorstep in 2007 asking about the cause of death, would they say he had been murdered? Or that he had been killed in an “accident”?
Ultimately, the answer to the question of how many Iraqis died as a result of the U.S. invasion will never be definitively known, and every attempt to provide one is bound to generate new methodological problems as well as political controversy. Still, the attempt must be made.
Survey methods of counting deaths while wars are still raging will always have major weaknesses and potential sources of bias. Human affairs continue to resist reduction, and certainty continues to elude the best efforts of social scientists. But this shouldn’t prevent us from giving best estimates of deaths in Iraq the same credence we give such estimates in Darfur and elsewhere. The best evidence we currently have suggests that the number of Iraqis who have died as a result of the U.S. invasion and the sectarian violence it unleashed is probably more than 400,000. A very high percentage of these deaths are civilian deaths.
The fate of the innocents of Iraq should be a matter of particularly sober introspection and moral concern among those Christians who defended the invasion as a “just war,” fully in harmony with Christ’s life and teachings and necessary in order to “preempt” a future terrorist attack on the scale of 9/11. A central tenet of the just-war tradition (and full disclosure requires that I make clear my own commitment to an ethic of Christian nonviolence in the Radical Reformation tradition) is that wars be strictly proportional, not causing greater harm than the harm they seek to prevent or redress. Another requirement is that wars be fought not merely for “the national interest” but always for the highest good of all people.
Purely in terms of lives lost, 400,000 Iraqi deaths would be 9/11 multiplied by a factor of 134. We should remember this when we grieve the 2,977 victims of the terrorist acts of 9/11 and the more than 4,400 U.S. soldiers killed in the war. These are individuals whose names we know with absolute certainty and whose lives we pledge not to forget. But for every one of them there exist dozens of Iraqis—many of them civilians—killed in an unjust war: fellow humans whose deaths we can hardly expect to remember if we do not agree to count them.
Related: A letter to the editor regarding this article & Osborn's response
The War in Iraq: How Catholic Conservatives Got It Wrong, by Peter Dula
Patrick Jordan's review of Anarchy and Apocalypse, by Ronald E. Osborn