Worst. Article. Ever.

Well, I haven't done an exhaustive study. But this article on counting casualties in Iraq is embarassingly bad, even by the standards of cursory reprints of press releases. As it happens, I've been writing about casualty counts in Iraq (so I'm afraid you're going to be hearing much more about this over the next month or so), and I'm flabbergasted by its bizarre omissions. I don't think that any credible person who has spent any time on the debate would be satisfied with this particular attempt to deal with it.

For starters, it repeats nearly uncritically the results of a survey from ORB which purports to find a million casualties in Iraq. Pretty much everyone I've spoken to regards their numbers--which utilize a murky methodology on a very small sample, and then publish bizarrely tiny uncertainty estimates on a survey in a bloody war zone--as a bit of silliness. It fails to mention that the largest survey using the best regarded methodology, which was just released by the WHO earlier this month, found about a tenth as many deaths.

But the really strange thing is that it drags in the first Lancet study without mentioning the second.

Medical journal The Lancet published a peer-reviewed report in 2004 stating that there had been 100,000 more deaths than would normally be expected since the March 2003 invasion, kicking off a storm of protest.

The widely watched Web site Iraq Body Count currently estimates that between 80,699 and 88,126 people have died in the conflict, although its methodology and figures have also been questioned by U.S. authorities and others.

ORB, a non-government-funded group founded in 1994, conducts research for the private, public and voluntary sectors.

It's as if he's unaware of Lancet II--which unawareness is really, really, really hard to achieve, because it's the main thing you get on either a google search or a Nexis of Iraqi casualties, or any plausible variant on those terms. Even Lancet I's authors would say that the study, which was published in 2004 on the cheap and had a very small sample, has been superceded by the larger, better funded, and more recent Lancet II, which found 601,027 violent excess deaths (654,965 total).