They suggest that the study over stated the deaths by as much as a factor of three, if I have read the report correctly.

Extensive? They manage to cover the big explosions and mass kidnappings, yes. But given that the vast majority of deaths are via gunshot, and given that media presence in Baghdad has decreased dramatically, no, I doesn’t give me pause. This is what’s going on in the background, 24/7.

I don’t know what bug crawled up Kaplan’s butt on this issue, but he’s been very, very sloppy. There were no mortality surveys conducted in Iraq between 1992 and the first Lancet study (there was one UN survey of childhood mortality in 1998 – conducted by one of the Lancet I authors). The “studies” he’s referring to comes from the boilerplate description of methodology for all countries. The UN estimate for Iraq was a pure guess, and he conveniently fails to mention two other guesses during that time period – the CIA’s was 5.3 and the U.S. Census Bureau’s 5.5.

One thing is true, though, the Lancet authors are indeed horrible at rebuttals. They should have torn Kaplan apart on his freshman-level Google-fu and the moronic, innumerate dartboard analogy he used in criticizing the first study.

But hey, Kaplan quoted the U.N. figures, not me…I didn’t double-check his work, but I’ve read his column for several years now, and he seems like a pretty reliable sort; I don’t imagine he just yanked that number out of his a…er, thin air…

However, according to data from the United Nations, based on surveys taken at the time, Iraq’s preinvasion mortality rate was 10 per 1,000.

Tell you what. If you can find these phantom surveys, or even a real reference to them, I’ll never darken your Lancet posts again.