"Not credible" was President Bush's quick verdict on the new study, published this week in the British medical journal the Lancet, calculating that more than 650,000 Iraqis have died as a result of the U.S. invasion and its ensuing chaos. It is understandable that the president would be quick to dismiss such an explosive claim, but the rest of us should take the time to look a bit more closely.
The number of estimated deaths claimed by the study is inconceivably huge and wildly out of scale with any previous figures we've heard. But it's difficult to avoid the conclusion that the human suffering in Iraq has been far beyond our imagining.
The peer-reviewed study's named authors include three researchers from the Bloomberg School of Public Health at Johns Hopkins University -- one of them is Gilbert Burnham, co-director of the school's Center for Refugee and Disaster Response -- and a professor from Baghdad's al-Mustansiriya University. Funding for the project was provided by MIT. These are not shabby credentials.
But academic degrees and prestigious affiliations alone do not establish truth. Bush said the problem is that the study's methodology has been discredited. But the team relied on a "cluster sample survey" technique that is frequently used for public health research, especially in the developing world.
No one should find the basic concept unfamiliar, since it underlies such mainstays of modern life as public opinion polls and market research. The survey team picked what was deemed a representative sample -- in this case, 1,849 households scattered throughout Iraq -- and used that sample to draw conclusions about the population as a whole. That's the same method pollsters employ to predict who will win an election.
Ideally, the selection of respondents should be as random as possible. The process of choosing the 50 widely scattered neighborhoods in which the Johns Hopkins team did its work was not quite ideal, but the Lancet peer reviewers who cleared the study for publication could find nothing that would significantly skew the results. Interviewers went house to house, recording detailed information about deaths before the 2003 invasion and deaths since.
The researchers tallied 82 pre-invasion and 547 post-invasion deaths in those households. The death rate per year nearly tripled after the invasion, they found, and a full 300 of the post-invasion deaths, or more than half, were the result of violence. (By contrast, only 2 percent of pre-invasion deaths were violent.) Of those killed by violent means, more than half died from gunshot wounds; the rest died mostly in bombings and airstrikes. Victims were primarily young and middle-aged men. In more than 90 percent of cases, family members were able to produce death certificates confirming what they told the interviewers.
Those may look like small numbers on which to base such large claims, but that's how survey research works. Pollsters in the United States, a much larger country, routinely predict nationwide trends on the basis of fewer interviews.
Does this prove, as the study asserts, that precisely 654,965 Iraqis have died "as a consequence of the war," and that exactly 601,027 of those deaths were due to violence? No, it doesn't. The Johns Hopkins team reports being 95 percent certain that the true figure lies between about 400,000 and about 900,000 -- a large range of uncertainty that some critics have seized upon as discrediting the whole project.
But the exact number is not the point. Rather, it's the scope and scale of the carnage.
Late last year President Bush gave an off-the-cuff estimate of 30,000 Iraqi civilian deaths -- this after the administration had steadfastly refused to acknowledge even trying to count the Iraqi dead. Now the administration is willing to allow that perhaps 50,000 civilians have died. It is unclear whether any science at all has gone into these estimates or whether they were essentially pulled out of a hat.
But quite a lot of science went into the Johns Hopkins study. Even if you assume that the number of Iraqi civilians killed since the war began is at the very low end of the study's range, that's still a quantum leap from earlier estimates. We now have reputable evidence -- not proof, I'll allow, but science-based evidence from respected scholars, published in one of the world's most prestigious medical journals -- that the humanitarian tragedy in Iraq is much, much worse than anyone had suspected.
If the study's findings are flawed, then its critics should demonstrate how and why. But no one should dismiss these shocking numbers without fully examining them. No one should want to.
Eugene Robinson