Shortly before the war in Iraq started, a small outfit named Iraq Body Count (IBC) coalesced in the ether of cyberspace with a singular goal: to tabulate the civilian deaths resulting from a U.S. invasion. The group predicted famines and plagues, citing U.N. surveys that said there would be “starvation and homelessness for millions,” and at least “two million refugees.” Within weeks IBC web counters, replete with the image of bombs dropping from a plane, peppered left-wing websites and noted the escalating civilian casualties, updated as soon as the now-massive IBC database was.
Two years later, those same counters are still omnipresent on the web, and they’ve just pushed past the 25,000 mark. But it’s a recently released IBC report parsing thousands of clips that’s getting all the mainstream media attention. Even though IBC is as partisan as they come, the media took the bait–hook, line, and sinker. And in the rush to publish the blaring headers of the report–U.S. forces killed four times as many civilians as “anti-occupation forces”!–hardly anyone examined the underlying data.
But they should. The report itself is premised on two years worth of newspaper and web data–well, “data” in a loose sense. IBC rests its laurels on numbers generated from newspaper reports of deaths and newspaper reports of mortuary and hospital logs. The methodology is flawed from the get-go, and though the citations are noted, there are no links to the articles. (That’s not exactly true; there are a couple links to particularly gory stories, like one with the headline, “I saw the heads of my two little girls come off.”) The lack of transparency, however, is only a small flaw in an ocean of methodological errors. Deaths only have to be verified by two of their accepted sources–which include (the non-fair and balanced) CommonDreams.org, Al Jazeera, and ReliefWeb–and often the second source is just a reprint of the first. A death count only has to be mentioned in passing in the article, like a doctor or bystander who gives a reporter casualty estimates.
IBC’s counters also give themselves wriggle room, by displaying a minimum and maximum body toll to account for situations where numbers may not be clear. These generally come from ambiguous interpretations of words like “probable” or “most”: For example, if a doctor says 50 people were killed in an air raid, and “most” were civilians, IBC will add 26 to its minimum, and 49 to its maximum. In other cases, they’ll invent percentages and play the same trick, by defining words as they see fit. Constant fiddling with numbers to generate minimums and maximums, and long-winded explanations of how this is done, provide the pseudo-scientific cover IBC relies on for its high-profile publicity.
Flaws like these are endemic, but specific machinations prove far more egregious. For the sake of a closer examination, the casualties can be divided into three roughly equal groups: deaths taken from mortuary records (the most outlandish), invasion-phase casualties, and then the rest.
A Statistical Morgue
The most disingenuous segment of IBC’s number padding is the addition of deaths from violence as reported by mortuaries. As of the two-year report in March, this included 8,913 people, over one third of IBC’s total. The rationale: Increased lawlessness as a result of the war has led to increased violence and therefore civilian deaths directly attributable to the invasion. IBC relies almost exclusively on media examinations, like this one from the AP. IBC first had to eliminate “background” deaths, those that would have occurred anyway and cannot be attributed to the war. So they took the AP numbers and subtracted similar violent-death numbers that were compiled by Saddam’s regime before the war; one suspects that Saddam Hussein, who once won 100 percent of the vote on 100-percent turnout, may not have been the most honest record-keeper. And of course, IBC excludes pre-war “violent deaths” like summary executions, or human-shredding machines. (In a note addressing this very point, IBC quotes Human Rights Watch as saying “the killing in Iraq [before the war] was not of the exceptional nature that would justify such intervention.”)
To head off anyone who might suggest that many of the mortuary deaths are actually insurgents, IBC cites one line from the aforementioned AP article in every explanation: “[T]he bodies of killed fighters from groups like the al-Mahdi Army are rarely taken to morgues.” There is no further information or substantiation in the actual article, and similarly, IBC doesn’t explain why it chooses four percent and two percent as the percentages to mark down the AP estimates, to take into consideration this “rarely” caveat. Blindly groping for numbers is a tactic IBC uses throughout, and despite tedious discourses, there is no scientific basis for any of the gerrymandering. But numbers like those from the mortuaries certainly help their bottom line.
Invasion Foibles
According to IBC’s numbers, 30 percent of civilians in the study were killed during the invasion, the period lasting until May 1, 2003. The largest incidents rely on the same manipulation as was done with the mortuary numbers. The first and highest death tally, which tops out with a 2,000 maximum and a 1,473 minimum, is based on an article from AP and one from Knight Ridder, both examining data from the 19 hospitals where dead civilians in Baghdad would have ended up during the period from March 20 to April 9. Even ignoring the incentive a Hussein-controlled hospital would have to inflate numbers, IBC twists and contorts itself to generate minimums and maximums. In this instance, the articles reported 1,101 civilian deaths and 1,255 “probable” ones. After subtractions to eliminate overlaps IBC may have already noted in its database, the question is: How to calculate “probable”? The answer: For the minimum, take half the number plus one; for the maximum, use the “bare minimum beyond certainty,” or the total number minus one. (See here for IBC’s full, tortuous (and torturous) explanation.) This may seem very technical, but it’s a bait and switch, a smokescreen to obscure a simple fact: The underlying numbers are worthless. Regardless of how you doll it up, and how much mathematical wizardry you apply, there’s no way to confirm these numbers, and there’s no way to divine what, precisely, “probable” really means in this context.
The second largest entry is a Los Angeles Times hospital survey, claiming a maximum of 978 and a minimum of 567–which overlaps with the time period for the above hospital survey. (The secondary source was CommonDreams.org, which only reprinted the L.A. Times piece.) Where these precise numbers come from is unknown, for the article mentions only that at least 1,700 were killed in Baghdad, according to 27 hospitals. In all likelihood, the same trick was used, whereby overlaps are subtracted, and then some percentage is calculated–rendering uneven numbers with an air of authenticity. The same L.A. Times article separately quotes an orthopedic surgeon south of Baghdad, who said there were more than 200 deaths at his hospital. Sure enough, IBC tosses in 200 in a separate entry, citing the same source. (At another hospital, in a different article, a doctor said a majority of the 400 he had treated were civilians; that entry has 201 dead.)
Next highest is an entry for 670 people killed in November, 2004 in Fallujah, and it is marked “provisional” and based solely on a U.N. report. IBC’s explanation of this entry is not enlightening. Then there are 633 deaths recorded in the initial battle for the southern city of Nassiriyah. But only one of the four cited sources could be found, and it only offered vague estimates from vague surveys, seemingly unrelated to Nassiriyah. The explanation for another battle in Fallujah in April 2004, for which IBC notes 616 deaths, is so convoluted that it is almost impossible to disentangle. See for yourself.
This laundry list is in no way meant to belabor the point. Rather, it is intended to show that very few of the largest entries–and the top 50 entries (of over 2,000) make up more than 50 percent of the total deaths–can be substantiated. All are second- or third-hand reporting with numerous sleights of hand. One gets the distinct impression that much of the commentary is only intended to launder nebulous reports into believable numbers.
The Remaining Third
So a third of the IBC death tally is from morgues and unsubstantiated, and a third is from the invasion and almost equally as unsound, which leaves somewhere in the 8,000 range. Many of these are smaller attacks, but a few more comments can be made on the overall methodology, and on some of the specific entries. First, IBC counts all “signed-up” army recruits as civilians, and the same goes for police. It’s not at all surprising, then, that in the two-year report, of the 2,280 people whose occupations were recorded, about 1,000 were police, and another 200 were in security fields. Members of the army who were first caught, and then killed, are considered civilians because “they had lost their capacity and status as combatants and could have been expected to be treated under the provisions of the Geneva Conventions.” When terrorists don’t abide by the Geneva Convention, chalk another one up for IBC. Another entry had an individual who died “apparently from natural causes” while being interrogated by the U.S. One “civilian” killed in the “service” industry was an “arms dealer.” There are two entries with 40 and 41 deaths as the maximums, and, amazingly, zero deaths as the minimum. The list of inconsistencies could go on and on.
None of this is to belittle the sobering fact that civilians have been and continue to be killed in Iraq. The problem, however, is that an organization with an obvious axe to grind is getting attention from numerous media organizations–and people are actually taking it seriously. Hard-and-fast numbers on civilian deaths would certainly be a boon to the national and international discourse on Iraq, but IBC is doing nothing more than blindly throwing darts at a dartboard. Unfortunately, the target this latest go-round was media attention and widespread disinformation. And Iraq Body Count hit the bull’s eye.
–Alston B. Ramsay is an associate editor of National Review.