Madam, - Edward Horgan writes (March 6th) of the "limitations" of the Iraq Body Count reports which estimate 60,500 post-invasion Iraqi civilian deaths to the end of 2006; but he ignores the glaring flaws in the methods of the Lancetresearch team which produced the figure of 655,000 deaths as of July 2006.
Mr Horgan has a point when he mentions the different methods used by the two studies. As it happens, however, we have a third report which used the same method - random household survey - employed by the Lancetteam to measure post-invasion mortality. In 2004, when the Lancetissued its first report, the Iraq Living Conditions Survey used a much larger and better-distributed sample (21,668 households compared with the Lancet's 990), with longer interview times (median 83 minutes per household compared with 20 or less) and larger and better-trained interview teams. The ILCS estimated 24,000 violent deaths, with a confidence interval of 18,000 to 29,000, a figure quite close to that calculated by Iraq Body Count. The Lancetestimate was 98,000 deaths, with a confidence interval stretching from 8,000 to 194,000, viz., a margin of error greater than 90 per cent! The latest Lancetfigure of 655,000 comes with a similarly wide margin of error: a confidence interval of 393,000 to 943,000. Readers can judge for themselves the relative credibility of these estimates.
Bio-statisticians have criticised the Lancet team's methods for the failure to allow for the possibility of unrepresentative clusters of households. They have voiced doubts about the randomness of the selection and the rigour of the interview process. There are also the common-sense objections. If 92 per cent of respondents could produce certificates for claimed deaths, why have fewer than 10 per cent of the 655,000 been officially recorded? Can we really believe that more than 98 per cent of heads of households or their spouses were at home and willing to participate when first called on? Violence capable of killing 655,000 would be expected to leave millions of wounded. Where are they?
The appearance of the Lancet's editor on anti-war platforms suggests that the journal is driven by a political rather than a scientific agenda. The 655,000 figure is now displayed daily on the news channel Al-Jazeera. Those intent on seeing the US as the fount of all evil in the world will continue eagerly to lap up inflated death tolls from such sources. Those who regard it, however flawed, as the chief bulwark against the global jihad will see them as the propaganda they are. - Yours, etc,
DERMOT MELEADY, Clontarf, Dublin 3.