The Number...654,965.....


If this had been on-line Eli would have passed it on to Tim Lambert who owns the Lancet Mortality Study franchise, but it came as an article in the Johns Hopkins Alumni Magazine (Feb 07, may be on line later) by Dale Keiger. Why Hopkins? Gilbert Burnham is co-Director of the Hopkins Center for Refugee and Disaster Response and Professor in the Bloomberg School of Public Health. Les Roberts is adjunct there and lecturer at the Columbia Mailman School of Public Health.

UPDATE: Now on line with comments at Deltoid (2/7/07).

There are several interesting nuggets in this article that Eli has not seen elsewhere. For example Roberts had approached Burnham in 2004 who agreed to collaborate and provide the funding.

In September 2004 Roberts packed $20,000 in his shoes and a money belt, and lying on the floor of an SUV, slipped into Iraq from Jordan to coordinate the data gathering. Over three weeks, six trained Iraqi volunteers surveyed 7,868 people in 33 locations throughout the country.

As far as the funding of the second study

Burnham and Roberts began pondering a second survey as soon as they had published the first. Roberts hoped someone else would do it and, as an independent source, verity their results. But in late 2005, a group from the Center for International Studies at the Massachusetts Institute of Technology approached them. They had $90,000 for research in conflict areas of the Middle East and sought advice on how to apply it. Burnham says, "We started discussing with them what you had to do. Within five minutes of starting that conversation, the MIR people said, "Would you guys like to do it?" The CRDR contributed additional funds, and Burnham, Roberts, and other Bloomberg School faculty set to work on a new mortality survey.

As to the design

Zeger (one of the collaborators on survey design), a professor of biostatistics, worried about "recall bias".....Zeger made sure that the second survey would cover not just the months elapsed since the conclusion of the first one, but go back over the period of the first study as well. If a new sample of people produced results closely corresponding to the earlier findings, that would go a long way toward validating both studies. He also advised taking a subset of 10 clusters from the first survey and sampling them again; not the same households, but the same neighborhoods.

Those who know more about the surveys would know if this was done. What about the accuracy of pre war reporting:

They (B&L) reminded people that even without the war, Iraqis should have been dying at the rate of at least 120,000 per year from natural causes, yet in 2002, before the invasion, the Iraqi government had reported only 40,000.

also

Raed Jarrar, a contributor to Foreign Policy.....He was country director of a 2003 door-to-door casualty survey, sponsored by the Campaign for Innocent Victims in Conflict (CIVIC), that counted 2,000 civilian deaths in Baghdad in roughly the first 100 days of the war. "I know how hard that [sort of survey is]," he said, and notes that his limited study recorded twice as many casualties as that period's official figure.

Fundamentally I agree with Burnham and Roberts that

the accuracy of our figures is not the most important aspect of this research. Yes, we are sure of the data. Yes, we are confident of our estimate. Of course the figures are important. But even if the number 654,965 is wrong, that's not the point. People whom the US has a duty to protect are dying. That's the point.

Johns Hopkins Magazine is slow getting stuff on line, but I will put up an update notice when the article appears.

UPDATE: Now accepting comments at Deltoid