17  |  Civilian casualties

221.  An Iraqi Government spokesperson commented on 12 October that “these figures

[in the study] are unrealistic and give a very exaggerated picture”.138

222.  A DFID Statistics Adviser provided advice on the study to IPU on 12 October, at

their request.139 He concluded that “in essence, the method is tried and tested”.

223.  Professor Anderson provided his views on the study the following day.140 He stated

that he had received comments on the study from an independent expert on statistical

epidemiology and demography. Professor Anderson advised:

“… the study design is robust and employs methods that are regarded as close to

‘best practice’ in this area, given the difficulties of data collection and verification in

the present circumstances in Iraq … The methods are an improvement on those

used in the 2004 Lancet article by the same author …”

224.  Professor Anderson advised that deaths were much more prevalent among

adolescent to middle‑aged men and suggested that bias might remain with respect to

the level of non‑combatant mortality.

225.  Professor Anderson concluded that, given the reasonably robust study design

and appropriate analysis methods, the UK Government should be cautious in publicly

criticising the Lancet study.

226.  An IPU official produced a brief on the study for Mr Blair later on 13 October.141

The brief summarised the advice from the DFID Statistics Adviser and

Professor Anderson, and identified several “key points”:

“• … People are dying at the hands of those who choose violence to pursue their

aims …

• There are conflicting estimates [of the number of civilian casualties] from a

number of sources, and no comprehensive or accurate figures;

• The numbers that the Lancet has extrapolated are a substantial leap from other

figures. There is no reason to assume the Lancet figure is any more accurate

than any other is.”

227.  The same IPU official wrote to colleagues on 16 October:

“… we do not (not) accept that the figures quoted in the Lancet survey are

accurate … The figures are extraordinarily high and significantly larger than the

figures quoted by the Iraq Body Count or Iraqi Government – however the survey

methodology used here cannot be rubbished, it is a tried and tested way of

138  Briefing IPU, 13 October 2006, ‘The Lancet: Iraq: a Cross‑Sectional Cluster Survey Sample’.

139  Email DFID [junior official] to FCO [junior official], 12 October 2006, ‘Foreign Secretary Comment on the

Lancet Report’.

140  Minute APS/CSA to DJC‑Sec10, 13 October 2006, ‘Iraq – Mortality After the 2003 Invasion of Iraq:

a Cross‑Sectional Cluster Sample Survey – Lancet October 2006’.

141  Briefing IPU, 13 October 2006, ‘The Lancet: Iraq: a Cross‑Sectional Cluster Survey Sample’.

209