Funerary Foibles
A correction: When the story about Iraq's new mortality estimate came out I was going to write something up, but then I saw the marvelous essay by Paul Craig Roberts and I knew I couldn't do as good a job, so I asked Craig if I could publish his here, which he very kindly gave permission to do. I agree 100% with Craig's sentiments but I have begun to have some sneaking doubts about that mortality estimate (which, I don't think, affects in any major way what Craig's saying). As some of you may know, I had an (in)famous article about mortality rates in Bosnia ("The Bosnia Calculation", New York Times Magazine, April 23, 1995). I took a lot of flack for that—a huge amount of flack—yet eventually my substantially lower, conservative estimate was proved much more accurate than the standard boilerplate tossed about by politicos and newshounds during and after the war. Let me make a similar argument here.
The first point is, one need not exaggerate the mortality in Iraq to know that what's happening is a terrible outrage against humanity. Having said that, let's think for a moment about what this new study claims: An average of approximately 500 deaths per day, due to violence, have gone unnoticed and unreported. Now, of course it's hard to know anything for sure in Iraq, even if you're there. But let's assume, knowing what we do know about Iraq, that most of these uncounted deaths have occurred in the center of the country where there's the most violence. (If they'd occurred in the North and/or South at that rate I don't think anybody would argue that they'd have gone unnoticed.) We know that any time there's mass killing both Sunni and Shia have a clear track record of reporting it. They hold press conferences, they distribute DVDs, they even give out the names and family phone numbers of the dead. Most visibly, they hold funerals. With killing proceeding at such a clip in central Iraq there'd be hysterical funerals every day, all over the place. But that hasn't been observed.
Now, I'm not familiar with the survey methodology used, and despite having read critique and rebuttal am not really competent to judge its theoretical accuracy. I'd rely more on what we know about the scene, which is essentially what I did for Bosnia. And although when I wrote up the Bosnia story I did talk with many, many people who knew about the situation—almost 100 sources, if I recall correctly—while here I have none such, it seems to me not to be unreasonable to map the same kinds of considerations onto the problem and thus find some general guidelines about testable hypotheses. In short, upon reflection I would urge caution about accepting these new mortality figures uncritically, or hanging too much of an argument on them.
Posted by George Kenney on October 19, 2006 9:04 PM 
« More Kudos to Frontline | Main | The Evening of Empire »
Comments
“This American Life” did an excellent show last year on the 2004 study, done by the same group and using the same sampling technique. That study had used a much smaller sample size, but came up with a similar figure (i.e. surprisingly high; an order of magnitude greater than the Iraq Body Count figure at the time). You can listen to the show here; it’s very interesting.
No matter what questions people might have about the accuracy of the findings, the President’s response to the study was certainly a delight. Now, his general erudition and impeccable knowledge of the classics have been widely reported on, but I had no idea he is also a formidable expert in statistics and research methodology. The man just never ceases to amaze!
:-)
Posted by: Jason | October 22, 2006 12:42 AM