Over a year ago, an international team of epidemiologists headed by Les Roberts of Johns Hopkins School of Public Health completed a "cluster sample survey" of civilian casualties in Iraq. Its findings contradicted central elements of the narrative of the war that politicians and journalists had presented to the American public and the world.

After excluding the results from Anbar province as a statistical anomaly and half the increase in infant mortality as possible "recall bias," they estimated that at least 98,000 Iraqi civilians had died in the previous 18 months as a direct result of the invasion and occupation of their country. They also found that violence had become the leading cause of death in Iraq during that period (51 percent or 24 percent with or without Anbar). However, their most significant finding was that the vast majority (79 percent) of violent deaths were caused by "coalition" forces using "helicopter gunships, rockets or other forms of aerial weaponry," and that almost half (48 percent) of these were children, with a median age of eight.

When the team's findings were published in the Lancet, the official journal of the British Medical Association, they caused quite a stir, and it seemed that the first step had been taken toward a realistic accounting of the human cost of the war. The authors made it clear that their results were approximate; they discussed the limitations of their methodology at length and emphasized that further research would be invaluable in giving a more precise picture.

A year later, we do not have a more precise picture. Soon after the study was published, American and British officials launched a concerted campaign to discredit its authors and marginalize their findings without seriously addressing the validity of their methods or presenting any evidence to challenge their conclusions. Today the continuing aerial bombardment of Iraq is still a dark secret to most Americans, and the media still present the same general picture of the war, focusing on what appear in the light of this study to be secondary sources of violence.

Les Roberts has been puzzled and disturbed by this response to his work, which stands in sharp contrast to the way the same governments responded to a similar study he led in the Democratic Republic of Congo in 2000. In that case, he reported that about 1.7 million people had died during 22 months of war, and as he says, "Tony Blair and Colin Powell quoted those results time and time again without any question as to the precision or validity." In fact, the U.N Security Council promptly called for the withdrawal of foreign armies from the Congo, and the U.S. State Department cited his study in announcing a grant of $10 million for humanitarian aid.

Roberts conducted a follow-up study in the Congo that raised the fatality estimate to 3 million, and Tony Blair cited that figure in his address to the 2001 Labor Party Conference. However, in December 2004, Blair dismissed the epidemiological team's work in Iraq, claiming that, "Figures from the Iraqi Ministry of Health, which are a survey from the hospitals there, are in our view the most accurate survey there is."

This statement by Blair is particularly interesting because the Iraqi Health Ministry reports, whose accuracy he praised, have in fact confirmed the Johns Hopkins team's conclusion that aerial attacks by "coalition" forces are the leading cause of civilian deaths. One such report was cited by Nancy Youssef in the Miami Herald on September 25, 2004, under the headline "U.S. Attacks, Not Insurgents, Blamed for Most Iraqi Deaths." The Health Ministry had been reporting civilian casualty figures based on reports from hospitals, as Mr. Blair said, but it was not until June 2004 that it began to differentiate between casualties inflicted by "coalition" forces and those from other causes. In the three months from June 10 to September 10, it counted 1,295 civilians killed by U.S. forces and their allies and 516 killed in "terrorist" operations. Health Ministry officials told Ms. Youssef that the "statistics captured only part of the death toll," and emphasized that aerial bombardment was largely responsible for the higher numbers of deaths caused by the "coalition." The overall breakdown (72 percent U.S.) is remarkably close to that attributed to aerial bombardment in the Lancet survey (79 percent).

BBC World Affairs Editor John Simpson reported on another Health Ministry report that covered the six months from July 1, 2004, to January 1, 2005. This report cited 2,041 civilians killed by U.S. and allied forces versus 1,233 by "insurgents" (only 62 percent U.S.). Then something strange but sadly predictable happened. The Iraqi Health Minister's office contacted the BBC and claimed in a convoluted and confusing statement that their figures had somehow been misrepresented; the BBC issued a retraction; and details of deaths caused by "coalition" forces have been notably absent from subsequent Health Ministry reports.

So, the British and American governments and the U.N. responded positively to Roberts' work in the Congo, and Iraqi Health Ministry reports support his findings in Iraq in spite of official efforts to suppress them. Official and media criticism of his work has focused on the size of his sample, 988 homes in 33 clusters distributed throughout the country, but other epidemiologists reject the notion that this is controversial.

Michael O'Toole, the director of the Center for International Health in Australia, says: "That's a classical sample size. I just don't see any evidence of significant exaggeration . . . If anything, the deaths may have been higher because what they are unable to do is survey families where everyone has died."

David Meddings, a medical officer with the Department of Injuries and Violence Prevention at the World Health Organization, said surveys of this kind always have uncertainty but "I don't think the authors ignored that or understated. Those cautions I don't believe should be applied any more or less stringently to a study that looks at a politically sensitive conflict than to a study that looks at a pill for heart disease."

Les Roberts himself has also compared his work in Iraq to other epidemiological studies: "In 1993, when the U.S. Centers for Disease Control randomly called 613 households in Milwaukee and concluded that 403,000 people had developed Cryptosporidium in the largest outbreak ever recorded in the developed world, no one said that 613 households was not a big enough sample. It is odd that the logic of epidemiology embraced by the press every day regarding new drugs or health risks somehow changes when the mechanism of death is their armed forces."

The campaign to discredit Les Roberts, the Johns Hopkins team and the Lancet employed the same methods that the U.S. and British governments have used consistently to protect their monopoly on "responsible" story telling about the war. By dismissing the study's findings out of hand, U.S. and British officials created the illusion that they were suspect or even politically motivated and discouraged the media from taking them seriously. This worked disturbingly well. Even opponents of the war continue to cite much lower figures for civilian casualties and innocently attribute the bulk of them to Iraqi resistance forces or "terrorists."

The figures most often cited for civilian casualties in Iraq are those collected by Iraqbodycount, but its figures are not intended as an estimate of total casualties. Its methodology is to count only those deaths that are reported by at least two "reputable" international media outlets in order to generate a minimum number that is more or less indisputable. Its authors know that thousands of deaths go unreported in their count, and say they cannot prevent the media misrepresenting their figures as an actual estimate of deaths. I have asked them several times to be more active in challenging such misrepresentations, but I have to acknowledge that the misrepresentations are so widespread that this would be quite a task.

Beyond the phony controversy regarding the methodology of the Lancet report, there is one genuine issue that really does cast doubt on its findings. This is the decision to exclude the cluster in Fallujah from its computations due to the much higher number of deaths that were reported there (even though the survey was completed before the widely reported assault on the city in November 2004). Roberts wrote in a letter to the Independent, "Please understand how extremely conservative we were: we did a survey estimating that 285,000 people have died due to the first 18 months of invasion and occupation and we reported it as at least 100,000."

The dilemma he faced was this: in the 33 clusters surveyed, 18 reported no violent deaths (including one in Sadr City), 14 other clusters reported a total of 21 violent deaths, and the Fallujah cluster alone reported 52 violent deaths. This last number is conservative in itself, because, as the report stated, "23 households of 52 visited were either temporarily or permanently abandoned. Neighbors interviewed described widespread death in most of the abandoned homes but could not give adequate details for inclusion in the survey."

Leaving aside this last factor, there were three possible interpretations of the results from Fallujah. The first, and indeed the one Roberts adopted, was that the team had randomly stumbled on a cluster of homes where the death toll was so high as to be totally unrepresentative and therefore not relevant to the survey. The second possibility was that this pattern among the 33 clusters, with most of the casualties falling in one cluster and many clusters reporting zero deaths, was in fact an accurate representation of the distribution of civilian casualties in Iraq under "precision" aerial bombardment. The third possibility is that the Fallujah cluster was atypical, but not sufficiently abnormal to warrant total exclusion from the study, so that the number of excess deaths was in fact somewhere between 100,000 and 285,000. Without further research, there is no way to determine which of these three possibilities is correct.

No new survey of civilians killed by "coalition" forces has been produced since the Health Ministry report last January, but there is strong evidence that the air war has intensified during this period. Independent journalists have described the continuing U.S. assault on Ramadi as "Fallujah in slow motion," devastating the city block by block. Smaller towns in Anbar province have been targets of air raids for the past several months, and towns in Diyala and Baghdad provinces have also been bombed. Seymour Hersh has covered the "underreported" air war in the New Yorker and writes that the current U.S. strategy is to embed U.S. Special Forces with Iraqi forces to call in U.S. air strikes as U.S. ground forces withdraw from Iraq, opening the way for heavier bombing with even less media scrutiny (if that is possible).

One ignored feature of the survey's results is the high number of civilian casualties reported in Fallujah in August 2004. It appears that U.S. forces took advantage of the media focus on Najaf at that time to conduct very heavy attacks against Fallujah. This is perhaps a clue to the strategy by which they have conducted much of the air war. The heaviest bombing and aerial assault at any given time is likely to be somewhere well over the horizon from any well-publicized U.S. military operation, possibly involving only small teams of Special Forces on the ground. But cynical military strategy does not let the media off the hook for their failure to find out what is really going on and tell the outside world about it. Iraqi and other Arab journalists can still travel through most of the country and news editors should pay close attention to their reports from areas that are too dangerous for Western reporters.

A second feature of the epidemiologists' findings that has not been sufficiently explored is the one suggested above by Michael O'Toole. Since their report establishes that aerial assault and bombardment is the leading cause of violent death in Iraq, and since a direct hit by a Mark 82 500-pound bomb will render most houses uninhabitable, any survey that disregards damaged, uninhabited houses is sure to underreport deaths. This should be taken into account by any follow-up studies.

Thanks to Les Roberts, his international team, Johns Hopkins School of Public Health and the editorial board of the Lancet, we have a clearer and very different picture of the violence taking place in Iraq than that presented by the "mainstream" media. Allowing for an additional 14 months of the air war and other violence since the publication of the Lancet report, we can now estimate that somewhere between 175,000 and 650,000 people have died as a direct result of the war; that 120,000 to 500,000 of them have been killed by "coalition" forces, and that 50,000 to 250,000 of these were children below the age of fifteen. In addition, the combined effect of conservative, even unrealistic, assumptions made to arrive at the lower of these figures makes it extremely unlikely that the actual numbers of deaths are close to the bottom of these ranges.

If you find yourself troubled or torn between accepting the "official story" of the war and the picture that emerges from the Lancet report, I would suggest the following. Both versions of events are efforts to tell a story or paint a picture from a patchwork of samples or snapshots taken in different parts of Iraq. However, the way that the samples are selected and pieced together is very different. In one case, the choice of samples and the way they are put together is clearly influenced and circumscribed by powerful political, military and commercial interests. In the other, the samples were chosen according to objectively established epidemiological practice, and the results were analyzed with scientific rigor.

As someone who has followed the reporting of this war very closely, I find the results of the study to be consistent with the picture that I have seen gradually emerging as the war has progressed, based upon the work of courageous reporters and glimpses through the looking glass as more and more cracks appear in the "official story." We are still left with civilian casualty figures that can only be described by very wide ranges. The responsibility for the failure to obtain more precise casualty figures and thus a more accurate view of this crisis falls fairly and squarely on the doorsteps of 1600 Pennsylvania Avenue in Washington and 10 Downing Street in London, two households that have experienced no excess deaths to children or adults as a result of the war.