- The Washington Times - Wednesday, January 4, 2006

The claim that 100,000 civilians were killed in Iraq derives from a study done by an international research team led by Les Roberts of Johns Hopkins University’s Bloomberg School of Public Health, which included researchers from Columbia University and Al-Mustansiriya University in Iraq. The research was done in Iraq in September 2004 and was published in the online edition of the Lancet on October 29, so that it would appear days before the November 2004 presidential election.

There is little doubt that this study was rushed into print to influence the 2004 presidential election. Roberts told the Chronicle of Higher Education that he hoped that the study would provoke “moral outrage.” His motivation was even clearer when he spoke to a Scottish newspaper: “I was opposed to the war and still think it was a bad idea,” he told the Glasgow Herald, “but I hope the science has transcended our objectives. As an American, I am really, really sorry to be reporting this.”

And the editor of the Lancet, a well-respected British medical journal, clearly shared Roberts’s political outlook. “The invasion of Iraq, the displacement of a cruel dictator, and the attempt to impose a liberal democracy by force have, by themselves, been insufficient to bring peace and security to the civilian population. Democratic imperialism has led to more deaths, not fewer,” said Richard Horton.

Later, Roberts claimed he had to rush the piece into print to protect the lives of Iraqi members of the survey team. Sure.

Still, politically motivated researchers can sometimes be right. The question is: how credible is the research? A closer look reveals so many confounding variables as to render the results provably false.

Roberts et al. conducted interviews at thirty sites, allegedly randomly selected and interviewed thirty-three people in each place. The survey teams did not ask for death certificates, trusting Iraqis to tell them who had died. “In the Iraqi culture,” they wrote “it was unlikely to fabricate deaths.”

But there is a cultural disconnect between Western researchers and Iraqi respondents, as Stephen Apfelroth of the Albert Einstein College of Medicine notes. The survey team asked how many deaths per “house-hold” were known to Iraqis. Americans think of a household as a nuclear family. Arabs think of a household as including the extended family, which can easily number in the hundreds. As a result, one hears about the deaths of distant in-laws and third cousins, even though many of these “relatives” might live far from the surveyed area.

Roberts et al. claim that their survey used random samples. But as Slate.com’s Fred Kaplan points out, the survey sites were not selected by chance. At times, the group couldn’t find enough people to interview, so they expanded the area they studied. Sometimes they couldn’t get access to a survey area because it was closed by checkpoints or war, so they went to another location that was convenient.

The Roberts study claimed a 99.5 percent response rate, meaning that virtually everyone contacted instantaneously agreed to participate in the survey. Veteran statisticians find this response rate surprisingly high. Far from being random surveys, it appears that Roberts’s team gravitated to the most vocal Iraqis.

Another bias is something statisticians call “cluster sampling.” Two-thirds of the deaths reported in the Roberts study were located in one cluster in Fallujah. Journalist Michael Fumento comments: “That’s it, game over, report worthless.” Fallujah, the scene of heavy fighting for more than two years, is simply not representative of Iraq as a whole.

Roberts and his co-authors conclude that “there were 98,000 extradeaths(95%CI 8,000?194,000) during the postwar period.” This means that, as Fred Kaplan writes, “that the authors are 95 percent confident that the war-caused deaths totaled some number between 8,000 and 194,000.”

At this point, any fair-minded observer would have to conclude that the Roberts study is pure “political science.”

A more responsible attempt to count civilian deaths in Iraq can be viewed at http://www.Iraqbodycount.net. Compiled by Hamit Dardagan and John Sloboda, the site counts civilian casualties, but rather than trusting Iraqis they meet in the street, Dardagan and Sloboda only count deaths if they appear in at least two online media sources. As of June 13, 2005, Iraqbodycount.net had estimated between 22,248 and 25,229 civilian casualties.

One unnecessary death is a tragedy; 100,000 phony deaths is a travesty.

Richard Miniter is the author of two New York Times bestselling books, “Losing bin Laden” and “Shadow War,” and is an internationally recognized expert on terrorism.



Click to Read More

Click to Hide