A startling new household survey of Iraqis released last week claims as many as 1.2 million people may have died because of the conflict in Iraq - apparently lending weight to a 2006 survey in the Lancet that reported similarly high levels.
More than one million deaths were already being suggested by anti-war campaigners, but such high counts have consistently been rejected by US and UK officials. The estimates, extrapolated from a sample of 1,461 adults around the country, were collected by a British polling agency, ORB, which asked Iraqis how many people living in their household had died as a result of the violence.
Previous estimates, most prominently collected by the Johns Hopkins Bloomberg School of Public Health, reported in the Lancet in October 2006, suggested almost half this number, 654,965, as a likely figure in a possible range of 390,000 to 940,000.
Although the household survey was carried out by a polling organization, rather than by epidemiological researchers operating under the discipline of scientific peer review, it has again raised the specter that the 2003 invasion of Iraq has caused a far more substantial death toll than officially acknowledged by the US or UK governments or the Iraqi Ministry of Health.
The ORB survey follows an earlier report by the organization which suggested one in four Iraqi adults had had a family member killed. Their latest survey suggests that in Baghdad that number is as high as one in two. The poll also questioned the surviving relatives on how their loved ones were killed. It reveals 48 per cent died from a gunshot wound, 20 per cent from the impact of a car bomb, 9 per cent from aerial bombardment, 6 per cent by accident and 6 per cent from another blast or ordnance.
If true, the latest figures would suggest the death toll in Iraq now exceeds that of the Rwandan genocide in which 800,000 died.
The new effort to estimate the number of dead in Iraq is certain to reignite the controversy over the lack of any proper accounting of the number of civilian dead in Iraq.