Madam, - Dermot Meleady and Sean Coleman have launched a strong attack on the study published by the Lancetwhich estimated the Iraq death toll at over 600,000 ("Counting the dead of Iraq", March 1st, 8th, 9th). However, this is only an echo of a debate which has been settled in the study's favour. Prof Gilbert Burnham and his associates at Johns Hopkins University in Baltimore and Al Mustansiriya University in Baghdad have quite effectively refuted all attacks on the methodology of the study.
The methodology in question, called "cluster sampling" is a tried and trusted technique used in public health studies all over the world. What is innovative is the use of an epidemiological tool to study war deaths. If the results have shaken the powers-that-be, well, maybe that is a good thing.
The Lancetis a prestigious, peer-reviewed scientific journal. This study was independently read and recommended by four experienced practitioners in the field. For the journal not to publish would have been a major dereliction of the editor's duty to publish scientific data and analysis without fear or favour. May I point out to Dermot Meleady that it is these values of free speech and free enquiry which he claims are being defended in Iraq?
Mr Meleady also confuses the technical terms "confidence interval" and "margin of error". The latter is defined in surveys like this one as half the confidence interval. The confidence interval contains the most probable number (654,965) and falls away each side down to the lesser numbers, where probabilities are almost zero. The colloquial term "margin of error" means "the real number could be anywhere in between", but in this study the statisticians are offering the number 654,965 as the most likely. The high (statistical) margin of error stems from the number of clusters in the sample. These were kept as small as possible to lessen the risk to the lives of the interviewers, all of them Iraqis.
Sean Coleman makes great play of the category "Deaths from unknown causes" in the report but he is being disingenuous. The meaning is clear enough from the context, and it means deaths where the family are uncertain of the perpetrators. This means that the killers could be coalition forces, insurgents, militia or the many criminal gangs operating in Iraq. Mr Meleady also offers the Iraq Living Conditions Study of 2004 as a convincing alternative. As its name suggests, this study was not concerned with mortality, and the associated questionnaire had only a single question on that topic. Its results are therefore quite dubious in this area, while being useful in other ways.
At the moment, the Lancetstudy holds the field as the only scientifically credible estimate of mortality since the Iraq war began in 2003. - Yours, etc,
TOBY JOYCE, Navan, Co Meath.