Cutting through the bullshit.

Wednesday 15 August 2007

'We do body counts'

Back in February, I made an attempt to project current estimates of excess Iraqi deaths since the march 2003 invasion on the basis of the estimates arrived at on the basis of the survey carried out by the Johns Hopkins University Bloomberg School of Public Health and al Mustansariya University School of Medicine team of Burnham et al between May and July 2006 and published in The lancet last October.

The Johns Hopkins survey report in October that estimated the number of excess deaths in Iraq since the March 2003 invasion at about 655,000 was conducted in May-July last year. Iraqis have obviously continued dying since then, and at an accelerating pace, from all reports.

Deploying a slightly suspect methodology, I have averaged the estimate, along with the upper and lower limits of the 95% confidence interval that the Lancet paper reported, to arrive at new estimates. On this basis, the probable number of excess Iraqi deaths as of this month stands at 769,583. Projecting from the July figures, we would be 95% certain that the true number lies in the range between 461,750 and 1,107,597. And counting. Since the numbers dying is on the increase, this is almost certainly an underestimation and the probability that at least 450,000 have died is therefore very high. Clearly there’s no point in even asking why these numbers don’t see the light of day in the mainstream press – only American deaths count.

So, of the estimated pre invasion population of 27 million, the occupation troops have already killed at least 450,000, and perhaps as many as 1.1 million – that would be over 4% of the population, and counting…

The approach I adopted was to calculate a monthly average for the first 40 months since the invasion covered by the Burnham et al study and assume that the average monthly rate of excess deaths for the year to July 2006 persisted. I also calculated on the basis of the average over the forty months, which yielded an unrealistically low projection, and on the basis that the monthly death rate post July 2007 increased by as much over the June 2005-June 2006 rate as the June 2005-June 2006 rate had increased over the average monthly rate May 2004-May 2005, which yielded projections that looked unrealistically high.

Using the same method and projecting out to this month, I arrive at an estimate of 1,015,195, including non ‘violent’ deaths.

A recent Information Clearing House linked me to the Just Foreign Policy site, where they have developed a new counter. The way it works is to update Burnham’s estimate of some 601,000 excess deaths specifically from violent causes over the period between the US invasion in March 2003 and the survey in May-July 2006 on the basis of increases in the Iraq Body Count counter. IBC provides minimum and maximum counts due to the range of numbers reported in the media they rely on. The new counter divides the average of the current minimum and maximum by the average of the two figures for July 2006 and multiply it by 601,000 to arrive at an approximation of the probable current death toll from violent causes.

JFP are quite explicit in acknowledging that this method is likely to under estimate the actual number of deaths because IBC relies on media reports. As the situation for reporters in Iraq has become more and more dangerous, the proportion of deaths reported in the media inevitably declines. They are also aware of the incompatibility between the IBC claim only to report the deaths of non combatants and Burnham’s principled approach to not attempting to distinguish combatants from non combatants.

A point that Eli Stephens has made and I believe I’ve mentioned before is that when the media report an incident that kills 5 and injures 30, that’s the end of the story. There is never a follow up story reporting on the 23 who subsequently died of their injuries. That is one source of the gross understatement of Iraqi deaths that IBC continues to defend.

JFP are also explicit that what their counter is trying to enumerate is the number killed specifically as a result of the US invasion and occupation since 2003. That means above and beyond the already elevated mortality from the previous US (aka UN) military adventure in Mesopotamia and the subsequent campaign of boycott and bombing.

I note with some pleasure that Tom Feeley at ICH has finally started presenting a more accurate description of the nature of the projection, ‘Number Of Iraqis Slaughtered Since The U.S. Invaded Iraq 1,004,219’. As I’ve mentioned before, until 18 June, ICH announced, ‘Number Of Iraqi Civilians Slaughtered In America's War On Iraq - At Least 655,000 + +’. There are two problems with this formulation. The survey on which the estimate is based explicitly made no pretense of distinguishing combatant from civilian deaths. And the breadth of the 95% confidence interval means that there is a small possibility that the actual number killed could have been as ‘few’ as 392,979 at the time of the survey, so it is not appropriate to describe the midpoint estimate as ‘at least’.

Anyway, I think the JFP methodology is sound and the projection it generates is not far off my 1,015,195 projection, perhaps a little less than the discrepancies regarding violent and non combatant deaths would suggest. So I’ve installed the code and will display the JFP counter here unless someone shows me why the reasoning is faulty.

No comments:

Post a Comment