Monday, October 16, 2006

Margin of Error

I'll come right out and admit it: I'm not a statistician, nor do I pretend to be. That said, I'm skeptical of the latest Lancet survey on deaths in Iraq.

In 2004, a month before the presidential election, a group from Johns Hopkins released a report saying that in the 17.8 months between the start of the Iraq war and September 2004, there had been 100,000 "excess deaths" in Iraq. That translates to an average of 5,618 deaths per month.

A second study was just published—again, notably, just a month before elections—that says there have been 655,000 deaths up to September 2006, or 555,000 in a 24-month period. That translates to 23,125 per month over the last two years. That number is around ten times higher than monthly totals that have been widely reported in the Western media.

My skepticism is not based in a facile trust of "official sources" or some misguided belief that the government and/or media wouldn't lie to us. I just don't see how the numbers add up. One would assume that the Iraqi death rate would be at its highest during the initial phase of the invasion, when coalition forces were bombing Baghdad and other Iraqi cities heavily. The second Lancet study suggests the opposite. In fact, according to these studies, the Iraqi death rate has essentially quadrupled during the long occupation.

Much criticism has been leveled at the methodology of the latest Lancet study. The figure of 655,000 is extrapolated from 629 verified deaths among a sample of 1,850 families consisting of 12,800 people. Again, I'm not a statistician, but this strikes me as a rather small focus group to arrive at such a large number. Granted, political polls use similarly small groups to predict the voting habits of the entire country, but voting trends and war casualties are completely different things. One would think the latter category is subject to a lot more variation.

My skepticism is buttressed by another study, also conducted on the ground in Iraq, which pegs the death toll at 48,343 through the end of September 2006. Someone is wildly wrong here. I hope, for humanitarian reasons, that it's the Lancet folks.

All deaths are tragic, and it doesn't really matter to a family that's lost a loved one if that death is one of 2,000 or 20,000 in a given month. Human emotions and statistics are notoriously bad bedfellows. Whether 50,000 or 500,000 Iraqis have died isn't necessarily the point. What matters is whether these people have died in vain because of the arrogance and criminal negligence of our elected officials. The morality of the war will not be measured primarily by the legions of the dead, but by the quality of the lives and futures of those who live.

We promised a bright and shining future to the people of Iraq. We've failed miserably so far on that bargain. If it's true that over 600,000 people have died in Iraq, that doesn't prove the failure, it just makes it much, much worse.
Listed on BlogShares