STATS ARTICLES 2004

2009 | 2008 | 2007 | 2006 | 2005 | 2004 | 2003


Counting Iraq's Dead
November 10, 2004
Rebecca Goldin
Confusion in reporting on deaths from the U.S. invasion.

The publication of a new study claiming that 100,000 Iraqis have died as a result of the invasion days before the American presidential election led many to wonder about its timing and accuracy.

The study, which appeared in the leading British medical journal, the Lancet, contained two analyses based on survey data collected by a group of interviewers in Iraq. Their goal was to ascertain the number of deaths before and after the invasion in order to gauge the impact of the war.

The first analysis included data from Falluja, the scene of some of the most intense fighting, while the second excluded data from Falluja, marking it as a “statistical outlier” (as the large number of deaths recorded there could skew the overall findings).

The analysis that excluded Falluja arrived at 98,000 “excess deaths” (i.e., deaths that would not have occurred if there hadn’t been a war), which the researchers described as a conservative estimate.

But while this figure was accurately reported, many media accounts missed an important distinction between the two surveys. For example, the figure of 98,000 deaths came from the survey that excluded data from Falluja; yet the study’s conclusion that Iraqis were “58 times more likely to die by violence” includes the deaths documented by the survey team in Falluja.

Similarly, the authors lump these two numbers together in their summary, as if they came from the same analysis:

"Making conservative assumptions, we think that about 100,000 excess deaths, or more have happened since the 2003 invasion of Iraq. Violence accounted for most of the excess deaths and air strikes from coalition forces accounted for most violent deaths."

Again, the first sentence is based on the data excluding Falluja, while the second is based on the entire data set, including Falluja.

This conflation is important given the authors admission that Falluja is a statistical outlier and should, perhaps, be left out of the study results - if only to err on the conservative side in estimating deaths.

The other problem in the way this study was reported is that some accounts attributed the 100,000 deaths to violence. Yet the overall figure includes nonviolent death associated with the war; these increased after the invasion due to accidents, heart attacks and strokes, infectious diseases, and neonatal or unexplained infant deaths.

According to the BBC, for example, “Poor planning, air strikes by coalition forces and a ‘climate of violence’ have led to more than 100,000 extra deaths in Iraq, scientists claim” (Iraq death toll 'soared post-war', Oct. 29, 2004). The National Post ( Canada) reported that “The rise in the death rate was mainly the result of violence, and much of it was caused by U.S. air strikes on towns and cities.” (Iraqi death rates soar since invasion: report: About 100,000 'excess deaths' mostly blamed on U.S. air strikes, Oct. 29, 2004).

Another source of imprecision in the coverage of the study was that of the 73 reported violent deaths 84 percent (all but 12) were due to actions of coalition forces. Yet, beyond Falluja, where 21 violent deaths occurred, the study does not report this percentage. In other words, we don’t know what proportion of violent Iraqi deaths is due to coalition violence.

The real message behind the numbers outside Falluja is that the tens of thousands of Iraqi deaths associated with this war are not directly due to military action. Reduced medical care, the stress associated with major changes, worsened infrastructure, and impediments to getting to hospitals can all contribute substantially to an increased death toll.

Regardless of where one stands on the current occupation of Iraq, this study reminds us that war is costly both because of violence extending into the civilian realm, and because war is associated with a nonviolent death toll as well.


Digg!

Technorati icon View the Technorati Link Cosmos for this entry