A recent study published in the highly respeceted British medical journal the Lancet, indicated recently that as many as 100,000 people have been killed in Iraq. Suprisingly, the US propaganda machine and the British foreign office have been muted over this. The reason? Because this figure comes from an almost indisputable statistical methodology. Even so, the figure of 100,000 has been mis-reported in many ways. This article seeks to explain the methodology and the actual figure that the study suggests is the real number of casualties is around 40,000.

The US led coalition decided beforethe war that they would record the number of civilian casualties killed in the ongoing war. This was a terrible and immoral error. Firstly it implied that Iraqi lives were worth less than coalition lives, where the death and wounding of every soldier was meticuolously recorded. Secondly, it allowed all sorts of speculation as to the true number of people that have been killed in the ongoing war - a gift for the anti-coalition forces. In a careful study however, even the wildest of estimates concerning the number of Iraqis killed appear to have been an underestimation.

A US based professor Les Roberts used a classic technique known as 'clustering' to try and establish a reliable estimate. 33 neighbourhoods were chosen at random, and then every household in that neighbourhood was interviewed. The number of deaths that had occured in those households from January 1 2002 including the date, cause and circumstance were recorded. They were then compared with the same data after the invasion on March 18th 2003, allowing the researchers to compare the pre and post invasion figures and reach a statistically sound number of Iraqi deaths as a direct consequence of the US led invasion.

The often quoted figure of 100,000 deaths however is misleading. This is the centre of the estimated range of deaths, the range of which varies from 8,000 194,000. The father one strays from the 100,000 mark, the less likely is the data to be reliable - 8,000 is as unlikely as 194,000. But using complicated statistical analysis on the data, there's a 90% probability that the actual death toll directly attributable to the US led invasion of Iraq is upwards of 40,000 people, with 60,000 people being the likely end figure.

Ironically, one of the randomly chosen clusters was Fallujah, and this was omitted from the final analysis in recognition that this might well skew the reaults upwards. The same statistical methodology has been accepted and used around the world for years, and the methodology is the rule rather than the exception and has tended to be validated rather than contradicted.

yechydda,