Civilian Casualties of War in Iraq

[Follow Ups] [Post Followup] [Our Discussion Forum]

Posted by Andreas from ( on Saturday, June 14, 2003 at 1:22PM :

Shlama all,

Yesterday Zinda - aka "The Weekly Under-Reporter" - ran a short article :

"Civilian Casualties of War in Iraq
11 June 2003
(ZNDA: Baghdad) The Associated Press has published a calculation of civilian casualties in the war based on records from 60 of Iraq's 124 hospitals. The estimate is that at least 3,240 people died, including 1,896 in Baghdad"

I enclose a more systematic study which unfortunately gives much higher tallies.




1) War may have killed 10,000 civilians, researchers say



War may have killed 10,000 civilians, researchers say

Simon Jeffery
Friday June 13, 2003
The Guardian

At least 5,000 civilians may have been killed during the invasion of Iraq,
an independent research group has claimed. As more evidence is collated, it
says, the figure could reach 10,000.
Iraq Body Count (IBC), a volunteer group of British and US academics and
researchers, compiled statistics on civilian casualties from media reports
and estimated that between 5,000 and 7,000 civilians died in the conflict.

Its latest report compares those figures with 14 other counts, most of them
taken in Iraq, which, it says, bear out its findings.

Researchers from several groups have visited hospitals and mortuaries in
Iraq and interviewed relatives of the dead; some are conducting surveys in
the main cities.

Three completed studies suggest that between 1,700 and 2,356 civilians died
in the battle for Baghdad alone.

John Sloboda, professor of psychology at Keele University and an IBC report
author, said the studies in Iraq backed up his group's figures. "One of the
things we have been criticised for is quoting journalists who are quoting
other people. But what we are now finding is that whenever the teams go into
Iraq and do a detailed check of the data we had through the press, not only
is our data accurate but [it is] often on the low side.

"The totality is now producing an unassailable sense that there were a hell
of a lot of civilian deaths in Iraq."

A spokesman for the Ministry of Defence said he had not seen anything to
substantiate the report's figures. "During the conflict we took great pains
to minimise casualties among civilians. We targeted [the] military. So it is
very difficult for us to give any guidance or credence to a set of figures
that suggest there was x number of civilian casualties."

IBC's total includes a figure of at least 3,240 civilian deaths published
this week by the Associated Press news agency, which was based on a survey
of 60 Iraqi hospitals from March 20 to April 20, when the fighting was
declining. But many other bodies were either buried quickly in line with
Islamic custom or lost under rubble.

Prof Sloboda said there was nothing in principle to stop a total count being
made using forensic science methods similar to those used to calculate the
death toll from the September 11 attack: it was a question of political will
and resources.

He said even an incomplete record of civilian deaths was worth compiling, to
assist in paying reparations and in assessing the claim before the war that
there would be few civilian casualties.

Lieutenant Colonel James Cassella, a US defence department spokesman, said
the Pentagon had not counted civilian deaths because its efforts had been
focused on defeating enemy forces rather than aiming at civilians.

He said that under international law the US was not liable to pay
compensation for "injuries or damage occurring during lawful combat

The Iraqi authorities estimated that 2,278 civilians died in the 1991 Gulf



Comment and Analysis

Earlier Analysis...

A Survey of Projects Counting Civilians Killed by the War in Iraq
June 12th 2003

Executive Summary

Overview and key conclusions

At the outset of the Iraq War, Iraq Body Count was providing the only systematic estimates of civilian casualties. Now, 15 different projects are at varying stages of completion. Press reports increasingly cite data from more than one project, and critical comparisons are becoming necessary.

This article critically reviews all projects that have made their existence known publicly, and summarises key project details in tabular form.

Taken together, the projects reinforce rather than contradict one another and provide converging evidence that current estimates putting the number of civilians killed at significantly above 5,000 are well-founded.

These projects have already informed immediate humanitarian efforts, and when complete, can feed into strategic considerations about the costs of modern warfare. Given the importance of estimating the civilian costs, it is surprising that no national or supra-national agency has yet contributed to this work.

Categorisation and comparison of individual projects

Projects differ on a number of dimensions. Two key dimensions are scope (comprehensive or limited) and sourcing (direct or indirect). Any given project must choose where to situate itself on such key dimensions, and accept the limitations that this imposes. Much published criticism of individual projects is misleading, because it makes the erroneous assumption that there is a single standard against which all projects can be judged. We argue that projects with differing aims and scope can complement and strengthen each other, provided they adhere to minimum standards of rigour and reporting clarity.

Four active projects, including IBC, are indirect projects that base their estimates entirely on published press and media reports. These projects differ primarily in their comprehensiveness and their handling of potential “double counting”.

Direct projects derive their data from on-the-ground investigations in Iraq. Four such projects, three of them based solely in the Baghdad area, have completed their work. The only one of these projects to have published a full and comprehensive report of its methodology and findings is a study by the Spanish Brigade Against the War, entitled “Evaluation of the attacks on the civilian population of Baghdad”.

The remaining projects reviewed are still in progress or have not reported any firm results as yet. Among these the direct project undertaken by the Campaign for Innocent Victims in Conflict (CIVIC) aims to be the most comprehensive, using 150 surveyors working in 11 major population centres around the country.


1. Introduction

Our primary role at Iraq Body Count (IBC) has always been to keep the civilian costs of the Iraq war firmly in the public eye. We have been able to maintain a consistent role in promoting public awareness through the adoption of a strict and openly-declared data-gathering methodology combined with innovative use of web technologies, particularly for the dissemination of our findings. This has involved the daily monitoring, sorting and compilation of media-reported civilian deaths in an expanding and open-ended public database, based solely on reliably recorded deaths (rather than on extrapolation or speculation), and the production of a “bottom-line” statistic giving a running total of the number of reported civilian deaths at every stage of the war and subsequent military occupation. The dissemination of this statistic has been greatly enhanced by the participation of thousands of independent websites, large and small, carrrying our real-time IBC Web Counters.

IBC’s work has also been a resource for the print and online media, with many individual articles citing IBC data (eg, Christian Science Monitor, 21st May 2003) and both the Guardian and Reuters making use of IBC figures to provide regular civilian casualty updates during various phases of the conflict.

We are encouraged that the aims and concerns underlying our work are receiving increasingly widespread acceptance, evidence of which comes from the growing number of projects that seek, in varying ways, to fulfil similar and/or complementary roles to ours.

A consequence of this positive development is that, increasingly, reporting about civilian casualties draws on a number of studies, and such reporting is often framed in the context of debates about what is the best, or most accurate, way to estimate the civilian cost of the war. These debates are vital and timely. This article is our own contribution to that debate. In it, we review, as comprehensively as possible, the attempts that have so far been made to estimate civilian deaths in a systematic way. We compare the strengths and weaknesses of these different attempts, and make some suggestions how the combined efforts of all those working in this area may be channelled and focused toward our shared aim of “establishing the truth of what happened,” an outcome which we believe will best serve the interests of the war’s victims, including the families and loved ones of those who were killed by it.


2. Distinctions between projects strengthen rather than weaken the enterprise.
There are two main distinctions that need to be made among the different projects which seek to estimate casualties:

1. Scope - comprehensive or limited.

“Comprehensive” projects aim to work towards providing an estimate of all civilian deaths caused by the Iraq War. “Limited” projects aim to provide an estimate which is deliberately limited in some way, either in timescale, geographical location, type of records or cause of deaths.

2. Sources - direct or indirect.

“Direct” projects obtain data from on-the-ground sources in Iraq (such as hospital or mortuary records, photographic and forensic evidence, and interviews with family and neighbours). “Indirect” projects rely on secondary sources (such as media and news agency reports, including published reports, usually in the form of summaries, of “direct” projects by NGOs and others).

Due to real-world limitations of time, money and human resources, there is an inherent conflict between these two approaches. The more direct a project is and the fuller and higher quality its evidence, the less practical it is for it also to be exhaustive and complete; and the more comprehensive a study aims to be, the less likely that it can also be as detailed and as fine-grained in the evidence it collects as a “direct” project.

This makes it pointless to ask which is “better”. Different types of project are needed. Direct projects can provide evidence of forensic quality (which would, for instance, be able to establish facts about the deaths of named individuals, the exact circumstances of the death, and where direct responsbility might lie for compensation or reparation). Indirect projects can draw on evidence from a range of sources to arrive at estimates of global cost or impact, which will be of importance to political, military, and strategic debates. Putting all the projects together can strengthen their total impact.

Another difference relates to the difficulties of achieving “total” completeness or “absolute” directness. These are unachievable, and possibly meaningless, concepts. Neither an indirect study such as ours, nor any “after-the-event” investigation will eventually arrive at a definitive and totally accurate count. A theoretically “perfect” count would require, at the very least, a totally accurate census of the Iraqi population on the eve of the war, and DNA profiling of each individual. It would also require the entire land area of Iraq to be covered with continuously running video recorders, whose data had all survived the war. However, the absence of such “perfect” data place limitations which are no different to the limitations that hamper any historical or legal investigation. Investigations must do the best they can on the basis of the available evidence, gradually improving the scope and reliability of the estimates. Combining different sources of evidence from different projects can strengthen the evidence base.

We detect, in some media discussions, a degree of confusion about the whole “counting the dead” agenda which often comes about by focusing on some particular shortcoming of individual projects. Any project may always be criticised for not meeting some standard it cannot achieve and never set out to achieve. It is always possible to criticise a direct project for not being complete, or an indirect project for lacking “scientific” standards of evidence. Any project can ultimately be criticised for not being “total” or “absolute” in the sense outlined in the previous paragraph. The cumulative effect of this type of comment is to help create in the public mind a general (and false) impression that we can know nothing reliable from the projects that have been undertaken to date, or might be undertaken in future, and therefore that the whole enterprise may be discounted. This conclusion may be convenient for some, but is not supported by the evidence.

A good example of media-generated confusion comes in a New York Times article filed by John R Broder from Qatar on 9th April, at the height of the slaughter of civilians in the battle for Baghdad. In this report a Central Command spokesperson was cited as saying that “the number of Iraqi dead were certainly high, but ultimately unknowable”. In the same report, Mark Burgess, a researcher at the Center for Defence Information in Washington (CDI - see below), was quoted as saying that various attempts to estimate casualties suffered from “dubious methodologies”. The article finishes by quoting Mr Burgess as saying “We just don't know and we might just as well make up a number”. What follows shows that there is a great deal that is knowable from the attempts so far undertaken.


3. The projects

All projects mentioned in this review are summarised in the table at the bottom of this page. Each project is named and assigned a table-linked number (#1, #2, etc.) reflecting order of mention in the text.

3.A. Indirect projects based on secondary sources of information

Iraq Body Count (#1 - IBC) began its work in January 2003, and its data base includes civilian deaths reported from January 1st 2003 onwards. The choice of January 1st for the start of the count (rather than March 19th) was deliberate. US/UK military actions against Iraq have been continuous from 1991 to the present day. The illegal US/UK patrols and bombing raids in the northern and southern “no-fly” zones have taken place on a weekly, and sometimes daily, basis until 15th March 2003. Between the 1991 Gulf War and early 2002 some 1,476 people had been killed by such bombing raids, according to official Iraqi figures.

Although not directly focused on civilian deaths, US Bombing Watch (#2), a project of the Colorado Campaign for Middle East Peace, has provided a unique research and public information resource. This project has recorded every single media-reported bombing raid carried out by US or UK warplanes in Iraq since January 1999 (and plans to provide data going all the way back to 1991). To date they have recorded over 300 bombing incidents between January 1999 and the March 2003 invasion.

IBC started counting civilian deaths from the beginning of the month in which it became active. The 2003 “body count” of civilians killed by US/UK military actions already stood at 15 on the eve of the March 19th invasion. At the time of writing, the IBC project consists of 126 separate database entries recording media-reported civilian deaths, amounting to a maximum (reported) number of 7,203 such casualties of the war on Iraq.

IBC’s methodology requires each report of civilian deaths entered into its database to have been published by at least two independent media sources meeting certain prescribed standards, as set out in our published Methodology, occasionally supplemented by findings from recognized NGOs such as the Mines Advisory Group (MAG - see below). A range of internal checks are also undertaken within the research team before a minimum and maximum number is published for each item in the data base. This methodology has been very useful in discounting inaccurate reporting (sometimes due to mistranscription of interview data), or initial estimates rushed into print in order to meet press deadlines. While the dual-publisher requirement has prevented some (otherwise impeccable) reports from being included in the IBC database, this has most often been only a temporary exclusion until other corroborating news stories became available. And because IBC always ties its entries to specific times, locations and other “identifiers” it has been able to deal with potential double-counting (for an example of this process in action, refer to the Notes to Incident x073 in our database).

Beyond evaluating whether a source is or was in any position to know, IBC has never passed judgement on the original source of the estimate. Original sources have been those which are normally available to journalists, including government-employed spokespersons of the parties to the conflict, eye-witnesses, relatives, doctors, hospital officials, mortuary and funeral directors, and workers or spokespersons for NGOs and aid agencies. In early phases of the war, some, though not all, news reports of civilian casualties relied solely on press briefings from Iraqi officials, whose estimated total had risen to 1,252 civilian deaths by the time of their final briefings on April 4th. Some of the Iraqi data was unusable in IBC’s database because it lacked the “identifiers” mentioned above, but in any event IBC’s current total from a variety of sources has risen since early April to a maximum of approximately 7,200 deaths, with only about 130 of these remaining solely derived from Iraqi press briefings. Despite early claims that their figures were inflated for propaganda purposes, it is now clear from the projects under review that the Iraqi figures were ultimately under-estimates, as are all those based only on information from hospitals, and were further hindered by the breakdown of Iraqi communication channels and the government's more pressing concern with its own survival. (A recent Associated Press investigation referenced below provides more detail on how the Iraqi statistics were produced.)

In this context it is surprising that Operation Iraqi Freedom Total Casualty Report (#3) includes only the Hussein regime’s composite estimate of April 4th in its web-based table (this was its “methodology” during the war as well - to simply repeat the most recent official Iraqi claim with a disclaimer). The authors of this project, sponsored by the Washington-based Center for Defence Information (CDI), state that “until reporters gain unfettered access throughout the whole of Iraq, it will continue to be our position that there is simply no way to clearly ascertain the true extent of Iraqi civilian casualties. When such a time does arrive, rest assured that we will give these figures the coverage that they deserve.” At the time of writing the last entry to this database appears to have been made on May 6th.

A more comprehensive and widely-sourced project is Casualties of the 2003 Iraq War (#4), which is the personal initiative of Pranav Jani, a teacher at Wagner College, New York. This project is based on what the author describes as “informal internet searches”.

At the time of writing this project lists 217 separate media-reported incidents in which there were casualties, either civilian or military, US or Iraqi, starting March 19th and going up to the present time. Unlike IBC this project makes no attempt to correct for possible double counting in differing reports (and is completely explicit about this). It provides data in the form of a 6-column table, listing Iraqi dead, Iraqi wounded, coalition dead, coalition wounded, and the dead and wounded of other nationalities. Civilian and military casualties are not separated out, but this is one of the few serious attempts to publish a comprehensive list of wounded as well as dead. Presumably because no mechanism is in place to prevent double counting in its “informal” coverage of media reports, Casualties of the 2003 Iraq War refrains from offering any totals.

The projects described so far are the only ones which have operated in the public eye from the very first day of the invasion of Iraq.

A promising but quickly aborted intervention was made by the Swiss Foreign Ministry's List of civilian victims of the war in Iraq (#5), one week into the war. On the last weekend of March the foreign minister Micheline Calmey-Rey was quoted as saying her ministry would publish a list of civilians killed in the Iraqi conflict and implied that Switzerland, as the depository state of the Geneva conventions, had a duty to draw attention to the innocent victims of the war. This was especially significant for being the first time such a project had been undertaken by a government, and it was apparently well-organized and already under way:

“Swiss Foreign Minister Micheline Calmy Rey Monday, March 31, was quoted by the German paper Frankfurter Allgemeine Zeitung as saying that the section - operational since Monday morning - gathers information and data from credible and documented sources, adding [that] names of new victims would be added on an hourly basis.

The initial data available so far reveals the dirtiness of U.S.-British warmongers, the fakeness of their claims about a clean war, as well as their indifference to the lives of innocent, unarmed Iraqi civilians, Rey, member of the ruling Socialist Democratic Party, was quoted by the paper as saying.” (IslamOnline)

However after a vociferous outcry from right-wing parties in the country the project was abandoned just two days after it was announced, with the Minister citing only the difficulty of finding reliable sources of information. It is perhaps telling that at the time most (though by no means all) of the reports of civilian casualties were being disseminated from the Iraqi Ministry of Information, and that one of the accusations being laid at her door was that the list of dead would “become part of the propaganda of this war.” No attempt was made by the Swiss authorities to re-activate their proposal as multiple and independent sources of information became increasingly available.

The only other project based on indirect data known to us is the Project on Defence Alternatives (PDA) report dated 21 May 2003 and entitled Civilian Casualties in the 2003 Iraq War: A Compendium of Accounts and Reports (#6). This Washington-based group collected excerpts of journalistic and other accounts, including those of the Iraq Peace Team (IPT - see below), into a compendium “meant to serve as a database for further investigation of the modes and dynamics of conflict that generate non-combatant casualties”. The authors write as follows:

“The accounts cover the period from 19 March through the middle of May 2003, but the compendium is by no means a complete and comprehensive accounting of civilian casualties in the war. That said, the compendium probably includes a majority of the accounts of major incidents of multiple noncombatant death known to have occurred up to 15 May 2003. Moreover, the mortality reports from hospitals and cemetery records recounted here, while not complete, are sufficient in scope to suggest the scale of collateral fatalities in the war.

The accounts derive almost exclusively from Western sources. None rely solely on official Iraqi government reports of casualties. The compendium also strongly favors accounts based on field investigations by journalists, eyewitness testimony, and the testimony of doctors, families of victims, residents of neighborhoods in which incidents occurred, aid workers, and cemetery personnel. The collection method included tracking numerous Western news sources on a daily basis and supplementing this with news database and Internet news searches. Notably, the effort focused on major cases of multiple civilian fatalities insofar as these (1) often generated more than one independent report and (2) constituted a subset of incidents that, while manageably small, nonetheless reflected a significant portion of the war’s cost in civilian casualties.”

The authors then strike the following note of caution:

“While the accounts collected here provide a basis for estimating the total number of civilian fatalities during the war, a useful estimate cannot be derived by simply adding together the death tolls given in the various accounts. Further analysis is required to address some of the inconsistencies in the accounts and to avoid the problem of double counting. The records of deaths from the Baghdad hospitals especially suggest the difficulty of separating civilian noncombatant deaths and civilian combatant ones. Such an analysis is attempted in a Project on Defense Alternatives memo accompanying this compendium.”

Unfortunately, at the time of writing (12th June) the analysis or “memo” referred to has not been made available on the PDA web site. Analyses designed to address such issues have been systematically developed by IBC (for an example see link from IBC Incident x073), and we await with interest PDA’s publication of their solution to the problems of varying accounts, double counting and distinguishing civilian from non-civilian deaths.

Useful features of the PDA report are its reasonably extensive direct quotes from key sources, and its organisation of most of its data in chronological order, graphically depicting the unfolding of the conflict. However, it appears to offer little fresh data that is not already included in Iraq Body Count (#1) or Casualties of the 2003 Iraq War (#4), and its explicit lack of completeness sits oddly with claims that its data is “sufficient in scope to suggest the scale of collateral fatalities in the war” or provides “a basis for estimating the total number of civilian fatalities during the war.”

We would like to sound our own note of caution on any assessment of total “collateral fatalities” which involves extrapolating from supposedly representative samples to the entire country (for which purpose PDA appears to be suggesting that its compendium is “sufficient in scope”). While this approach might have some limited value under circumstances where evidence could not otherwise be expected to emerge, there is no reason to assume this about a fairly well-developed and well-educated society such as Iraq’s; nor are we convinced that such extrapolations could have the same value as, let alone substitute for or improve upon, those conducted in the more painstaking manner of counting those deaths that have actually been recorded - and IBC has accordingly resisted the temptation to make such extrapolations based on its own much larger and arguably more “representative” database.


3.B. Projects based on direct research carried out on the ground in Iraq

To our knowledge, four direct research projects have been completed and published up to the present time, and another three have been announced as under way but not completed.

3.B.1 Completed direct reports (in chronological order of publication)

Three of the completed reports have focused on civilian casualites in Baghdad, and there is probably considerable overlap between them.

On 26 April a group of researchers associated with the Spanish Brigade Against the War’s Arab Cause Solidarity Campaign published a report entitled Evaluation of the attacks on the civilian population of Baghdad (#7), which

“[R]elates 42 documented cases of attacks on the Iraqi civilian population carried out by the Anglo-American forces in the metropolitan area of Baghdad between 20 March and 5 April 2003, mostly by aerial bombing and missile strikes, but also land-based attacks that took place during the initial phase of occupation of the city. These cases have been documented on the ground by the Spanish Brigades group that was present in the Iraqi capital from the beginning of the war until 9 April, when US troops entered the area of the city where we were staying. Also included are reports of nine hospital visits, with testimonies of attacks that have not been sufficiently identified.”

The report includes a very detailed account of the methodology employed. The core data consists of 114 questionnaires completed by surviving victims of these attacks or their relatives who were direct witnesses. These were augmented by records from 5 hospitals covering the entire area of Baghdad: the Al-Kindi Hospital, the Al-Yarnouk University Hospital, the Saddam Hospital Complex; the Al-Nourman Hospital; and the Saddam City Hospital. Interviews were habitually conducted within a few hours of the attack victim’s admission, but were followed up during more prolonged interviews with the affected families, usually in relatives' houses. This immediate and direct information was complemented by data from visits to the places bombed and interviews with the neighbours of affected families.

Reconstruction of the stories of the families of bombing victims was laborious and occasionally fruitless. Due to the systematic destruction by US/UK bombing of the telecommunications system, and the impossibility of summoning ambulances to collect the victims, when an attack took place, families and neighbours would take injured and dying to different hospitals. As these were also unable to communicate with one another, they could not inform family members about their respective admissions.

For these and other reasons, the authors caution that the information in the report cannot be considered to be an exhaustive record of all coalition attacks (and consequent casualties). Rather “it should be regarded as a significant report of their breadth, systematic nature and severity”. However, the data compiled in their report document a total of 204 civilian fatalities and a further 583 injuries, making an average of 4.5 dead and 13 injured per attack.

The authors conclude, on the basis of what is undoubtedly the most direct and painstaking analysis of casualties so far to emerge from the war that

“the damage caused to the civilian population during the three weeks in which Baghdad was attacked was in no way due to mistakes, nor did it represent the “collateral damage” of a tactical surgical war, whose sole objective had been the destruction of the city’s governmental and military infrastructure. Our opinion.. is that the attacks were premeditated, designed to cause the greatest possible number of civilian victims, many being carried out repeatedly against densely populated and poor areas of the Iraqi capital. The logic of this conduct can only be explained by the deliberate will of the American and British political and military leaders to provoke terror and undermine the resistance of the Baghdad population.”

This study has not been reported by a single English-language news or media agency, to our knowledge.

Although not providing cumulative estimates of numbers of civilian casualties, the work of the Iraq Peace Team (IPT - an initiative of the Chicago-based Voices in the Wilderness organisation) is worth mentioning in this context. On April 4th this group published a report entitled “Civilian casualties and infrastructure damage in the 2003 US-led attack on Baghdad March 20 - April 1 2003”. This report is based on first hand observation and on interviews conducted by the Iraq Peace Team in Baghdad hospitals and neighbourhoods. Given the very limited resources of the team and given the constraints of operating in a war zone, the authors acknowledge that their report is a mere sampling and an unsystematic one at that. However, the Spanish Brigades Project has used some of the data provided by the Iraq Peace Team in its own more systematic report, as has the indirect project by the PDA (#6). One advantage of the IPT report not shared by any other we deal with is the inclusion of a considerable amount of photographic evidence of damage to individuals and to buildings. It is also the earliest compilation of civilian casualties to come out of Baghdad, but because it does not include estimates of numbers dead, it is not included in our table as a “project”. Much like the Spanish Brigade’s project this group’s work has been largely ignored by the mainstream media, with only the news website providing regular updates on their activites.

On May 4th, three journalists working for US-based Knight Ridder Newspapers published a report on the Death toll in 19 Baghdad hospitals (#8). Various stories providing differing degrees of detail about this study were widely syndicated through the Knight Ridder stable of newspapers, which operate in most US states.

No full account of the research has been published, but what can be pieced together from these press reports is that the estimates were “gleaned from archives that separated military from civilians” and “included those killed between March 19, when the US air war began, and April 9 when the city fell to American forces”. Although Iraqi doctors interviewed admit that their records were not perfect (war conditions did not allow for that), the data supplied were claimed to be accurate. Only seven of the 19 hospitals in the study are mentioned by name, although the overwhelming majority of the dead were said to have been counted at just three, named hospitals:

“The records show 1,101 deaths that doctors felt were clearly those of civilians, 845 of which were recorded at three hospitals - Al Kharama, Al Askan and Yarmuk - near the Baghdad airport.

An additional 1,255 dead probably were civilians, doctors say, all reported at the same three hospitals near the airport. At Al Kharama, 30 per cent of 450 such bodies belonged to women and children, doctors said.

Others were men without identification in civilian clothes who the doctors believed were civilians. But a final determination was not made, in part because of the enormous volume of bodies to be dealt with.”

The same records show more than 6,800 wounded. The report also provides evidence that the biggest number of deaths occurred on April 5th and 6th when US troops began fighting their way into the city. This report suggests that the Spanish Brigade researchers were able to take witness evidence relating to fewer than 10% of the total civilian deaths caused by coalition action in the battle for Baghdad.

On May 18th the Los Angeles Times published a report by its staff writer Laura King entitled Baghdad’s Death Toll Assessed (#9). The headline to this story reads:

“At least 1,700 Iraqi civilians died and more than 8,000 were injured in the battle for the Iraq capital, according to a Los Angeles Times survey of records from 27 hospitals in the capital and its outlying districts. In addition, undocumented civilian deaths in Baghdad number at least into the hundreds and could reach 1,000, according to Islamic burial societies and humanitarian groups that are trying to trace those missing in the conflict.”

Like the Knight Ridder project (#8) there is no full report of the research beyond that which is reported in the single LA Times story on the work. Similarly, only a minority of the hospitals surveyed are mentioned by name in the press report. The LA Times project records casualties up until 24th April, some two weeks after the fall of the Iraqi regime. This means that some fatalities from unexploded ordnance are included in this project, as well as from continued fighting after 9th April. Also notable is that the study is not restricted to hospital records but includes mosque-based volunteer burial associations who dealt with“undocumented” deaths, often of those who died on stretches of road.

The failure of either the Knight Ridder or the LA Times projects to give an exhaustive compilation of deaths by hospital means that it is almost impossible to know how much overlap there is between these two projects, or how either of them relate to the Spanish Brigade project, even though all three projects centre their work on Bahgdad hospitals during the same period of time.

For meaningful use to be made of these two newspaper-funded projects it is essential that a full and rigorous compilation of data is either published or made available to a competent authority to combine with, and compare to, other data. This brings us to the most recent and ambitious press study, published on June 10th:

“Someone has taped together the shredded binding, as if that could fix the horrors inside. There are pages bathed in dried, reddish-brown blood, their letters smeared and unintelligible. The frantic scribblings and bloody handprints are a record of war.”

This vivid pointer to the conditions prevailing in Iraq’s hospitals forms the introduction to another hospitals-based investigation conducted over five weeks by a team of Associated Press (AP) journalists. Covering civilian deaths recorded in the month between March 20 and April 20, the Associated Press Tally (#10) was restricted to 60 “accessible” hospitals (out of 124) throughout Iraq, and concluded that these hospitals recorded “at least” 3,240 civilians killed by the war.

Though geographically more wide-ranging than the other newspaper tallies, AP is refreshingly open about its incompleteness. Stating that “the count is still fragmentary,” AP notes that if ever there is a complete toll, it “is sure to be significantly higher” :

“Even if hospital records were complete, they would not tell the full story. Many of the dead were never taken to hospitals, either buried quickly by their families in accordance with Islamic custom, or lost under rubble.

The AP excluded all counts done by hospitals whose written records did not distinguish between civilian and military dead, which means hundreds, possibly thousands, of victims in Iraq’s largest cities and most intense battles weren't reflected in the count.”

Another point in the AP study’s favour is its publication of a breakdown of most of its findings by hospital in “key” cities:

• Baghdad: 1,896 (recorded at 24 hospitals)

• Najaf: 293 (four hospitals)

• Karbala: 200 (one hospital)

• Mosul: 118 (five hospitals)

• Samawa: 112 (two hospitals)

• Nasiriyah: 145 (three hospitals)

• Fallujah: 89 (one hospital)

• Madain: 71 (one hospital)

• Diwaniya: 61 (one hospital)

• Kut: 52 (two hospitals)

• Tikrit: 45 (one hospital)

This is useful, but it would have been better still for each hospital to be named. Another (self-imposed) limitation is that the AP study attempted to “protect the integrity of its tally” by excluding numbers not based on daily record-keeping:

“In Basra, Iraq’s second biggest city, hospitals issued 413 death certificates and officials estimated 85 percent were for civilians, but the hospitals did not keep daily records listing civilian or military status of casualties. The AP did not include any Basra deaths in its count.”

This means that the grand total of civilian deaths in Basra for the AP study is zero (instead of around 350, if it had trusted “officials’” - actually, doctors’ - estimates), giving good grounds for the authors' conclusion that the actual numbers, both in Basra and elsewhere, were significantly higher.

Neither the Knight Ridder, LA Times or AP studies are defined by any specific aims. Presumably they are the results of normal journalistic enterprise, including the delivery of an exclusive story which can be used to the benefit of the employer and the career of the journalist. It would, therefore, hardly be expected that these journalists would directly co-operate with one another. Yet such co-operation would have ensured that there was no duplication of effort, and that meaningful summative compilations could have drawn on these reports.

3.B.2 Projects still under way, who have not issued final reports

The best publicised project currently under way is the Survey of civilian deaths in Iraq (#11) conducted under the auspices of the Campaign for Innocent Victims in Conflict (CIVIC), led by Marla Ruzicka, an independent US peace activist (most recently reported by ABC News on May 28th). She has organised 150 surveyors who are travelling the length and breadth of Iraq, conducting interviews door-to-door with victims and witnesses in the worst-affected areas.

At the time of the ABC interview this project has documented 620 civilian deaths in Baghdad, 256 in Najaf, 425 in Karbala and as many as 1,100 in Nasiriyah, and researchers were still at work in Kirkuk, Najaf, Ramadi, Amara, Kut, Diwaniyah, and Basrah. This makes the project potentially the most comprehensive and labour intensive project so far in Iraq. In this context it is surprising that, according to press reports, the work currently receives no public or NGO funding, and that funds are down to the last $50 (though donations may be made via their website).

Ruzicka came to public attention in 2001 for a somewhat similar project undertaken in Afghanistan, using local researchers. Her work convinced Senator Patrick Leahy (Democrat- Vermont) to insert language in an appropriations bill allocating $3.75 million to help Afghan victims (but of which sadly nothing so far has been disbursed, according to the CSM). Leahy has been responsible for ensuring that the Iraq War Supplemental Bill (signed by President Bush on 16th April 2003) directs that an unspecified portion of the $2.4 billion appropriated for relief and reconstruction in Iraq should pay for “assistance for families of innocent Iraqi civilians who suffer losses as a result of military operations”. Ruzicka’s concern for getting money to victims is admirable, but it places on her a particular responsibility to provide well-documented primary data. Although the CIVIC website says that “CIVIC will share survey results with all NGO’s who are interested in conducting similar studies in Iraq”, the only reports of this work currently available are rather cursory press reports. This limits the current usefulness of this study. We hope that a full and detailed report of the research will be made available by CIVIC as soon as time and resources permit

A second project announced as under way is the Survey of dead, injured, and missing (#12) by the Iraqi Red Crescent. The most recent report of this work was given in a Christian Science Monitor article dated May 22nd. In that article, Haidar Taie, head of the department for tracing missing persons, is reported as saying “thousands are dead, thousands are missing, thousands are captured”. Another official told the CSM that “In Baghdad we have discovered 1,000 graves, and that is not the final figure. Every day we discover more”. At one time it was reported that the Red Crescent would be issuing total figures by mid-May. It is now nearly mid-June and no report has appeared. In the same CSM report, two researchers from two different teams who “asked not to be identified until the evidence was clearer” were quoted as saying the number of civilian deaths might be as high as 10,000. Until we know who these researchers are, and what methodology they are using, we must, however, treat this estimate with caution.

Focusing on mines and abandoned or “dud” munitions, a Survey of deaths through unexploded ordnance in Northern Iraq (#13) has been reported undertaken by the UK-based Mines Advisory Group (MAG), and on a smaller scale by the US-based Human Rights Watch (HRW). MAG is also involved in co-ordinating removal of unexploded ordnance in various locations, as well as public education projects to try to avoid further death and injury. Cursory and undetailed reports appear on the websites of both organisations, and have been variously reported in the press. Both groups claim large numbers of continuing deaths after the end of the hostilities. On 27th April, HRW claimed that the number of civilians killed or wounded since the war ended in northern Iraq was higher than it was during the conflict. On April 17th MAG published a statement saying that

“Fifty-two killed and 63 injured by landmines and unexploded ordnance (UXO); these are the devastating statistics in just one week coming from the main hospital in Kirkuk, northern Iraq. The real figures will be much higher as we know that many deaths are not recorded. There is no death registration system.”

There is no more recent report from either organisation than end April, and no clear indication of scope or methodology of data-collection. There is at least one press report that HRW is currently conducting a general survey of civilian casualties in Iraq, but no announcement, details or mention of this are on HRW’s website, and it has therefore not been added to our table below.

3.B.3 Statements of intent (projects planned but not underway)

Our round-up of projects concludes with mention of two projects which have been announced, but appear not to have published any outcomes at all, even interim ones. The first, and apparently more comprehensive project is that announced on 29th April by the charity MedAct entitled The short, medium and long-term health effects of war on Iraq (#14). This project plans to bring together a whole range of health indicators, including, of most direct relevance to our concerns, direct and indirect deaths (combatants and civilians by age and gender on both sides) and deaths from post-war violence. Involving a very impressive list of international advisors, the project promises that its “full overview of the short- and medium-term effects of the war will be published in the autumn.”

The final project is referred to in connection with the Iraq war in an April 15th Washington Post article. A study on civilian suffering in war (#15) is being conducted under the auspices of the Carr Center for Human Rights Policy at Harvard University. The director of this project, Sarah Sewall, is quoted as saying that obtaining totals would be an unrealistic aim “given the size, intensity and speed of the US campaign” but that “investigating at least some incidents would not only bolster US credibility but also contribute to better military planning next time by understading the actual effects of particular US battlefield decisions”. There is as yet no information published on scope, sources, methodology or a delivery date for this study.


4. Summary and conclusions

15 projects focused on estimating civilian deaths have been reviewed, taking a variety of approaches, differing in scope, methodology, and degree of completeness. Although those that have reported outcomes tend to report somewhat different totals, this is entirely understandable as a function of the different scope, methodology, and completeness of each respective project. The important thing to note is that there is very little overall inconsistency. Nothing in one project contradicts or casts doubt on data from any other project. All projects must be viewed as incomplete “snapshots” of different grain. They generally reinforce rather than contradict one another. None so far has ventured to extrapolate from limited evidence to a (dubious) conclusion about the entire conflict. All recognize that more evidence remains to be gathered, but based on what is so far known, they provide converging evidence that the total number of civilians killed in the conflict is well above 5,000.

The fact that the various researchers are largely independent of one another, and have initiated their work without detailed collaboration or co-ordination, can be seen as both an advantage and a disadvantage. The advantage is that the total enterprise does not fail if any one project fails. The disadvantage is, of course, dissipation of resource and duplication of effort.

Considerations that are often felt to be of importance in evaluating projects of this sort are (a) the track record of the researchers; and (b) their political orientation. Our view is that these considerations are secondary. The primary criterion on which projects of this sort should be evaluated is the quality and degree of disclosure of their published materials (methodology, rationale, sources). The more explicit, detailed and public the project, the less relevant the background of the researcher becomes to any judgement of the merit of the work. In general, bias that is openly stated and recognized is preferable to that which is hidden in self-professed, and in truth perhaps unattainable, “neutrality”. However, what ultimately matters is the quality of the work.

Our main plea to some of our fellow researchers is, therefore, to publish as much detail as possible, including all sources consulted, the methodologies used, and a breakdown of deaths according to place and date. We also consider it good practice to publish the names and brief vitae of all main project personnel, together with an email address to which comments and questions can be directed.

We ask the press and media to assist this process by providing full sourcing for projects they mention, including contact details. [snip - maximum size exceeded]

-- Andreas
-- signature .

Follow Ups:

Post a Followup

E-Mail: ( default )
Optional Link ( default )
Optional Image Link ( default )

This board is powered by the Mr. Fong Device from