Prague, 31 October 2005 (RFE/RL) -- The sound of explosions can be heard almost daily in the Iraqi capital, Baghdad.
Since U.S. troops entered Iraq in March 2003, Washington has kept close tabs on how many U.S. soldiers have died in such explosions or other violence, as well as in noncombat accidents in Iraq. That count now stands at over 2,000 U.S. soldiers.
But until recently the Pentagon has been reluctant to try to estimate the number of Iraqis who have died or been injured due to fighting.
As retired General Tommy Franks, the former head of the U.S. Central Command responsible for troops in Iraq, once told reporters, "We don’t do body counts."
But now that policy appears to have changed.
The Pentagon reported to the U.S. Congress this month that some 26,000 Iraqis have been killed or wounded in attacks by insurgents since the start of 2004. The victims are both civilians and members of Iraqi security forces.
Western news agencies reported yesterday that the statistics appeared in a graph used to illustrate a quarterly audit of Iraqi operations.
Some U.S. analysts say the new Pentagon policy responds to U.S. domestic pressure to show that Washington cares about both U.S. and Iraqi casualties.
Anthony Cordesman, a military expert at the Center for Strategic and International Studies in Washington, D.C., says Pentagon officials have "begun to realize that when you focus on the U.S. it gives the impression that the U.S. doesn’t care about Iraqis."
Hard To Count
But are the figures accurate?
Moayed al-Haidari, Radio Free Iraq’s bureau chief in Baghdad, says getting accurate figures for civilian casualties is almost impossible. "One kind of problem with accurately counting the number of the people killed or injured in daily terrorist incidents is that the people who lose some member of their family, some of them are afraid to call the police and to go through the procedures for [registering] the case. Because some of those people are afraid of revenge from the terrorists, sometimes," al-Haidari says.
Al-Haidari says that in some cases, the police even refuse to go to areas to investigate killings by suspected terrorists, because they fear the officers themselves will come under attack.
And, out of fear of reprisals, some people also do not register their dead with hospitals but proceed directly to burial.
Given such uncertainties about the death toll in Iraq, the Pentagon itself has said its estimates are just that -- estimates.
The British daily "The Guardian" on 31 October quotes a Pentagon spokesman, Greg Hicks, as saying "The [U.S.] Defense Department doesn’t maintain a comprehensive or authoritative count of Iraqi casualties."
Still, the figures released this month are dramatically lower than some other estimates of Iraqi dead and wounded.
The independent U.S.-British organization Iraqi Body Count puts just the number of Iraqi civilians killed since the U.S. military intervention in 2003 at between 26,700 and 30,100.
Iraqi Body Count compiles its figures from media reports.
And in October last year, the British medical journal "The Lancet" published a study stating that some 98,000 Iraqis had died since the start of the war due to violence and other war-related conditions.
The study was compiled by epidemiologists from Johns Hopkins University in the U.S. city of Baltimore. It estimated the number of war-related deaths by comparing Iraq’s current death rate with the number of mortalities that occurred in pre-war years.
The Johns Hopkins University study -- now a year old -- continues to be widely disputed, especially by the British and American governments.
Still, analysts say that whatever the true number of Iraqi deaths and casualties, all the studies agree on one thing.
And that is that the number of Iraqi casualties is growing -- and it is civilians, not soldiers, who are the most frequent victims.
Among the latest attacks, a suicide truck bombing killed at least 30 people and injured dozens more on 29 October.
The bombing occurred in the small central Iraqi town of Howaider, near the provincial capital of Al-Baquba.