There's quite a lot to be said analyzing, interpreting, even advocating about the 'surge' and what we should be doing or not doing in Iraq. But in the last few days my main attention has gone to trying to pick apart the bamboozling statistics we're getting on the one hand and to find some reliable ones on the other. And like one of the intel analysts quoted yesterday in the Post, I find that once you pull away the first level of transparently bamboozling numbers you find that there simply aren't any reliable statistics at all.
The whole debate is a mix of apples and oranges comparisons, numbers tossed around whose provenance is never clearly explained and so forth. And the folks with guns on the ground in Iraq -- the US military, the Iraqi 'government', etc. -- who are best placed to compile numbers are actively taking steps to conceal them.
The one set of numbers we've found that appears to go back some way (a couple years) and have a consistent methodology are those compiled by the Associated Press from police reports about deaths in Iraq. To further the confusion, though, the AP seems unwilling to assemble these numbers together in one place, so you need to go back and piece together the separate monthly numbers from individual stories.
So far, with some sleuthing yesterday by myself and Spencer Ackerman, we've got these numbers.
Jan 07: 1,604
Feb 07: 1,552
March 07: 1,572*
May 07: 2155
June 07: 1640
July 07: 1760
August 07: 1809
To be clear, we don't think there's something magical about the AP's methodology. I can imagine various problems with measuring civilian death rates from police reports. The key is that appears to be a reasonable methodology and they appear to have used the same methodology going back to mid-2005. So these numbers give us at least some starting basis to measure change over time.
Here's where we need your help. There's a real needle in a haystack quality to finding these AP statistics. You can see the ones we have above. If you'd like to help, we're trying to track down the numbers for 2006. If you can find them, send us an email with the subject "Iraq Numbers". In the email include the month, the number, the citation of the article where you found the number and the sentence -- word for word -- in which the number appears. If you have a link to an article online, all the better. If it looks like we've got one of the numbers for 07 wrong, by all means let us know that too.
We're also working on investigating a couple other sets of numbers, so if you know of some, we'd like help on that front too. Again, we're looking for sets of numbers that give some view into what's happening in Iraq. We're looking for ones that go back for some period of time into the past and use a consistent methodology.