TCS Daily

Casualties of the Press

By Iain Murray - March 4, 2002 12:00 AM

No war in history, with the possible exception of the 30-minute conflict between Zanzibar and the British Empire, has been conducted without civilian casualties. Modern public opinion, however, at least in the West, demands that those casualties be kept to a minimum. Some have suggested American bombs killed 4,000 innocent civilians in Afghanistan. Now evidence is emerging that those figures are exaggerated. Do we need to hold our military to account?

Despite Pentagon denials of individual incidents, there is no doubt that, on many occasions during the war, bombs went astray, hit non-military targets or otherwise killed innocents. However, there is equally no doubt that the Taliban, conscious of the Western desire to minimize innocent casualties, deliberately exaggerated the number of civilian dead in its propaganda. This has now been confirmed beyond contest by the Afghan journalists themselves. They told the Associated Press on Feb. 12 that "Taliban officials systematically doctored reports of civilian deaths ... in an attempt to galvanize opposition to the bombing."

Any estimate that relies on stories that gave credence to news releases from the Taliban's mouthpiece, the Afghan Islamic Press (AIP), is therefore irrevocably tainted with this propaganda. Unfortunately, the most widely quoted figure for civilian deaths (about 4,000) fell into this very trap. Professor Marc Herold of the University of New Hampshire compiled a detailed list of news reports of civilian casualties from global media outlets. In including Pakistani and Indian news media he thereby included a large number of reports that drew from AIP sources, fatally weakening his argument.
That, however, is not the only problem with Professor Herold's analysis. His count suffers from numerous other inaccuracies such as:

Double-counting. A cluster bomb incident in Shekar Qala, near Heart, killed 8 people plus one who later picked up an unexploded bomb. Herold seems to count this twice, on both Oct. 22 and Oct. 25. The deaths of Gul Ahmed, his wife, six children and his neighbor's two children seem to be counted three times each (twice on the same day). There are several other examples.

Confusion of status. Herold counts as civilians at least eight Afghans slain in Landi Khiel near Tora Bora on Dec. 2. The reports make clear, however, that these were allied mujah'addin, who should not be included in a tally of civilian casualties, even if they were killed by mistake. There are also several examples of people claimed to have been killed in attacks in what were almost certainly legitimate targets, such as the compound of Mullah Omar or a Taliban-controlled power station.

Simple inaccuracy. Mohammed Pardis is listed as having been killed on Oct. 17. The Chicago Tribune article cited in support, however, actually quotes Pardis saying his house was destroyed that day. There are other, similar mistakes throughout the report.

Professor Herold's data selection and treatment are therefore both open to question. A STATS review of the data suggests that, on a careful reading, only 650 of the deaths he claims are in any way reliably reported. Giving the benefit of the doubt in some cases, such as those where western reporters were told that more people had been killed than they saw graves, raises the total to about 1100.

These numbers fit very well with two more circumspect reviews announced recently. The Associated Press have examined hospital records, visited sites and interviewed eyewitnesses and have come up with an initial estimate "in the mid-hundreds," although they consider it likely that that figure will rise. Carl Conetta, of the Massachusetts based Project on Defense Alternatives, performed a similar press review to Professor Herold, but deliberately kept to reputable sources, and estimates between 1000 and 1300 civilian casualties. These numbers seem much more soundly based than the Herold figures.

There is, however, little to be pleased about even if the figures are much lower than first thought. STATS' review of the stories Professor Herold cites suggests that, even on a conservative estimate, several hundred innocent civilians were killed during the battle around Tora Bora and the other mountain fortresses. If, as now seems likely, the senior Al Qaeda leadership were not there during the battle, the civilian death toll associated with it begins to look disproportionate.

It is a truism that the first casualty of war is truth. That is precisely why, when investigating the conduct of war, we owe it to ourselves to protect truth as completely as we can. In this case we both deserve and require reliable numbers. Now it seems that we are getting those, it is time for the Defense Department to address the questions they raise.

TCS Daily Archives