It was an October surprise courtesy of the Lancet medical journal. A report, rushed to the public via online publication five days before the 2004 election, claimed the American-led coalition had directly or indirectly killed about 100,000 Iraqi civilians since the invasion — most from airstrikes. The media, with no great love for Bush and already turning against the war, went wild.
The Lancet was so delighted with the reaction (if not the “wrong” election outcome) that in 2006 it updated its figure to a stunning 655,000 deaths. Further, this time it said violence directly caused all deaths. This paper, by amazing coincidence, appeared just before the mid-term election.
There were critics, including yours truly. But now there’s even more ammunition in the form of a statistical analysis by David Kane presented at the Joint Statistical Meetings in Salt Lake City. Naturally Kane’s assessment is under vicious attack not by proponents of good epidemiology but rather opponents of the war, primarily a troll at the website Deltoid, Tim Lambert.
(Normally a “troll” is someone who gets his jollies through specious attacks on others blogs. In Lambert’s case, he began his own blog to give him wider range and even alters individuals’ Wikipedia entries.)
To come up with its 2004 figure, lead author Les Roberts of the Johns Hopkins Bloomberg School of Public Health in Baltimore and fellow researchers sent Iraqis to interview 998 families in 33 neighborhoods across the country. They asked how many people in each household had died and of what, then extrapolated to Iraq as a whole. Thence the 100,000 figure, which they insisted was “conservative.”
BUT CONSIDER JUST THIS: Because the sample size was so small, the range for deaths was wider than Rosie O’Donnell’s rump: 8,000 to 194,000. So the Lancet researchers merely split the difference. They said the tiny sample size was necessary because the interviewers were in constant danger — assuredly from being caught in the crosshairs of an F-15 Strike Eagle.
Further, the researchers used death certificates but didn’t feel bound by them — interviews were fine. “In the Iraqi culture it was unlikely for respondents to fabricate deaths,” they wrote. Sorry guys, but I’ve reported from Iraq three times and I’ve written that interviewing Iraqis is essentially worthless because “they just tell you what they think will prove advantageous to them.”
Aside from the timing of the papers’ release, and bearing in mind that motivation alone is not grounds for conviction, consider that Roberts admitted to the Associated Press in 2004, “I was opposed to the war and I still think that the war was a bad idea,” adding, “As an American, I am really, really sorry to be reporting this.” Yeah. Right. Sure.
Thank goodness the tradition of disinterestedness in medical journals continues.
WANT MORE EVIDENCE the researchers knew their paper wasn’t worthy of lining bird cages? The 100,000 figure is allegedly the excess over pre-war Iraqi mortality, which they claimed was 5.0 per 1,000 people annually. That was a fabrication absolute vital to the overall calculation. According to the CIA World Factbook, the pre-invasion (2002) rate was over 20 percent higher at 6.07 per 1,000. Remember, the study was allegedly looking for excess mortality; therefore the lower the authors set pre-war mortality the higher the excess post-invasion mortality looks.
Consider, too, that 100,000 deaths during the survey period meant an average of over 180 a day, of which the Lancet attributed a majority to airstrikes. Have you heard anyone claim our airstries killed over 90 civilians on any one day during the entire course of the war?
Anti-war and anti-American groups even said the Lancet figure was ridiculous. The website iraqbodycount.org estimated at the time about 14,000-16,000 deaths since the war began, a figure that even now ranges from about 68,000 to 74,500. The Evil One himself, bin Laden, in his pre-election video, made reference to the Iraq war and stated “over 15,000 of our people have been killed.”
Lancet supporters like Lambert ignored all of this, for obvious reasons. But they’ve, well, raised Kane over the Kane paper.
THE FOURTEEN-PAGE PAPER by Kane, a fellow at Harvard University’s Institute for Quantitative Social Science, is so complex that unless you have a strong background in statistics it may cause your head to explode. But it contains one conclusion that, aside from all of the above information, renders the Lancet study worthless.
The Lancet in 2004 stated, “More than a third of reported [post-invasion] deaths and two thirds of violent deaths happened in the Fallujah cluster. This extreme statistical outlier has created a very broad confidence estimate around the mortality measure and is cause for concern about the precision of the overall finding.” An “outlier” is a figure that is so numerically distant from the rest of the data that it’s often treated as an error and hence rejected for use in the final assessment.
But Kane decided to include it for a very good reason. If two-thirds of the death data come from Fallujah, how can it possibly be considered an outlier? Politically inconvenient, maybe, but not an outlier. Upon including the Fallujah data, Kane concluded that as incredibly wide as the confidence interval was as given, it now became so wide “that the lower bound is negative.” This means the figures the Lancet came up with are not of statistical significance and therefore, in statistician lingo, “don’t count for squat.”
Kane does grant that some unknown figure in the full dataset just might alter his conclusion. But keep in mind that Kane took the Lancet data at face value; therefore he didn’t factor in the fabricated too-low mortality rate. Nothing in that dataset is going to alter his conclusion. In any event, the Lancet authors refuse to release their figures. Now folks, this isn’t like KFC protecting its secret recipe of 11 herbs and spices. Releasing the data would make everything transparent and settle this once and for all, which is exactly why the Lancet and the authors keep it under armed guard on a remote desert island.
Lambert, in his blog, after three years of blasting critics of the Lancet study, conceded, “I would suggest that [Kane] has proven that this confidence interval is wrong,” indeed, “obviously so.” Then he went back to bashing those same critics (self included) while letting his acolytes do the heavy lifting in his comments section.
Kane patiently provided over 50 responses to those comments. For the most part, the critics pretend to be dispassionate defenders of proper statistical analysis. But occasionally a comment appears that really speaks for them such as, “David, thanks for your answer. In other words, mass carnage and slaughter. A vast crime against humanity, for which the occupiers are obliged to pay reparations.” Or: “David, you’re just window-dressing a problem . . . . There are clearly two types of town [sic] in Iraq — those being blown to shit by Americans, and those being only slightly blown to shit by Americans.”
Kane was also repeatedly lambasted for allowing conservative blogger Michelle Malkin to post his paper, notwithstanding that he did the same for leftist blogger Tim Lambert.
Finally, Kane realized that responding to anybody who reads and comments on Deltoid is spitting into the wind. It’s enough that he’s done a great service in further exposing the truth about civilian casualties in Iraq and exposing a once-great medical journal that has abandoned accuracy for advocacy.
Michael Fumento, a health, science, and military writer, has been embedded three times in Iraq and once in Afghanistan.