Are Civilian Deaths in Iraq Rising or Not?

The counting of Iraqi civilian casualties is a hotly politicized topic in Washington these days. Depending on one's vantage point, the numbers are either sloping upward or downward. Generals in Iraq say the numbers of deaths are down. Anti-war Democrats say they are up, or at least, aren't budging much from their late-2006 highs. It also depends on which data is used and how it's interpreted (average monthly killed versus average daily killed). Employing the latter, the Pentagon's most recent quarterly report [PDF] refutes notions that the surge has resulted in a drop in violence, given that civilian casualties daily were lower in the few months preceding the surge than they were afterward. But interpreted differently, as General David Petraeus did before Congress in September, there has been a 45 percent drop in total fatalities against civilians, including noted reductions in sectarian bloodshed, IEDs, and "high profile" attacks. So which version of events or data is correct?

Here's my take on the topic: First, it is extremely difficult to gauge with any accuracy how many civilians are killed in Iraq, so any empirical findings should be greeted with skepticism. That is because census data is poor; many deaths go unreported; and there is often double-counting because it is oftentimes difficult to determine a civilian from a combatant. Morgue, hospital or government statistics are also notoriously unreliable. And any statistical analysis by outside researchers generally relies on cluster sampling, which poses problematic issues with sample sizes because of Iraq's shifting demographics.

Second, even if casualties are down overall this year, they are still dangerously high and way above 2004-2005 levels, not to mention that whatever number is reached does not include those tortured or kidnapped. Plus, many Iraqi Muslims simply bury their dead and do not go through an undertaker or hospital.

Third, trend lines are often short-lived and futile to predict. Proclaiming civilian casualties are sloping downward is akin to saying the insurgency is in its last throes. Plus, these trend lines obscure the fact that any temporary decline in casualties can be explained because of the ethnic cleansing of once heterogeneous Iraqi cities and neighborhoods. Of course, an Iraq comprising just Sunni Arabs could result in zero sectarian casualties but a homogeneous Iraq is presumably not the end goal.

Finally, the fault for all this confusion lies with the U.S. military, which decided early on it was not worth counting Iraqi civilian casualties. Even the term given to their tragic loss at the hands of U.S. forces -- "collateral damage" -- smacks of bureaucratic hubris. (A chilling segment on 60 Minutes reports that 30 civilians killed was the magic number Pentagon officials could live with when targeting a "high-value" terrorist in Afghanistan; anything higher requires approval from the defense secretary or president.)

The Pentagon should keep closer tabs on Iraqi (and Afghan) civilian deaths, especially when the United States is at fault. Only the American military has the presence on the ground and resources to catalog reliable statistics. It won't, for obvious PR reasons, but it has in the past. After military operations in Somalia and Kosovo in the 1990s, for instance, the Center for Disease Control and Prevention conducted body count estimates of local civilians. They should do the same for Iraq and Afghanistan.

Comments

[criscokid]

There is a source that has been documenting Iraqi civilian deaths by violence. See
www.iraqbodycount.org

[RumiSouth]

This is the standard Republican response to unwelcome data: stop keeping records, stop keeping count. Make the problem go away by ignoring it.

[wldnswmmr]

"And any statistical analysis by outside researchers generally relies on cluster sampling, which poses problematic issues with sample sizes because of Iraq's shifting demographics."

Thus, another offhand dismissal of the Lancet study, a peer reviewed statistical analysis carried out by Johns Hopkins University and financed by the Massachusetts Institute of Technology. The sampling technique used was the "gold standard" for such studies, and the results reached gave the normal 95% confidence level, with a small margin of error. Why is this study given such short shrift? It echoes Bush's unlettered opinion that the study was "methodologically flawed." In what ways, specifically? Statistical inductive reasoning is used all the time for policy decisions. Why is this particular study singled out for casual dismissal? The study does not take long to read, and its methods are uncontroversial. Is it simply that the results are so horrifying?

[Judd16]

The numbers are staggeringly good. But don't let that get you down, I'm sure there will be some more deaths for you to get excited about soon enough.