Since April 2005, the AP has tracked Iraqi casualty data by relying on hospital, police and military officials and morgue workers, as well as reporters and photographers at the scenes and verifiable, two-source witness accounts. I tracked down the data that the AP had already reported, compiled it month by month, and the AP reconfirmed it for me to ensure that I didn't neglect or misunderstand something. Comparing AP's figures for each month in 2007 to those for its 2006 antecedent -- see the first chart -- 2007 is more deadly for Iraqis than 2006 was. Looking just at the post-February 2007 surge, the numbers for Iraqi casualties fluctuate month by month, and show no clear decline in civilian casualties.
The second list we obtained is from Iraq Body Count, whose methodology centers on "cross-checked media reports, hospital, morgue, NGO and official figures." Its numbers tell a more complicated story. Unlike with the AP's figures, the IBC's reflect a minimum-maximum range owing to questions about the exact numbers of civilians killed, for instance, in a particular bombing, or the civilian status of others. (See the second chart.) But they generally find that the surge hasn't resulted in significantly lower civilian casualties for Iraqis this year. So far, the first four months following the surge are deadlier overall than their 2006 counterparts, with May 2007 being about as bad as May 2006, and June 2007 being better than June 2006.
IBC's Hamit Dardagan, to whom I am indebted for breaking these numbers down by month and explaining IBC's methodology, caveats the organization's assessment by saying that the more recent data are incomplete. The figures after March 2007, a month into the surge, are "probably lacking hundreds (though never into the thousands) of deaths each month, which it would contain if these months were as 'finished' as the earlier periods." IBC is continually revising earlier months' figures as new data becomes available. So expect the numbers to be revised upward in the coming months.
If you chart the month-to-month civilian casualty statistics just in 2007 then the post-surge monthlies in the AP's count toggle a bit, but they hit a high in May, drop significantly in June, and creep back upward afterward. In the IBC's count, since the surge begins, the numbers drop in April, tick up slightly in May and then drop in June. That pattern roughly tracks with the AP's findings, giving confidence in the methodology of each. (However, the post-April 2007 figures are probably going to be revised upward.)
What the figures do show, however, is that 2007 remains more deadly for Iraqis, month for month, than did 2006. The two exceptions, May and June 2007 in IBC's count, may not stay exceptions for long, but they count as less deadly months for Iraqis than the previous year. In neither count did Iraq experience fewer than 1000 civilian casualties each month in 2007.
Neither of these counts include other important metrics, such as population displacement or sectarian killings; nor killings by Shiite death squads vs. Sunni insurgents; or incidences of car bombings or other mass-casualty events.
Finally, let me extend my personal thanks to the literally dozens of readers who wrote in with invaluable research help. You're what sets this medium apart. From what I've read over the last several days, I know I can trust you to correct me with any mistakes.