The Dance of Death ... to the tune of 654,965 dead Iraqis ...

The Dance of Death (painted by Sudeth Kumar)

At a news conference today, President Bush said (in reference to the roughly 655,000 figure), "I don't consider it a credible report. Neither does General (George) Casey (top commander in Iraq) and neither do Iraqi officials [our puppets]."

I wonder how many of these people have grappled with the problem of counting the dead from disasters like the tsunami in Sumatra or the 1906 San Francisco fire or ... the Holocaust? If they had, they'd understand the need and the accuracy of "cluster sampling." President Bush would rather hear the numbers given to him by his "yes" men, just as Donald Rumsfeld wanted to only hear numbers given to him by his "yes" men even before the invasion when obtaining estimates of necessary troops to invade, win and hold (forget rebuilding!) Iraq.

At least his expert Pentagon officials refused to comment on the efficacy of the numbers -- or the techniques used in deriving them, knowing full well that "silence is golden" in the vicinity of President George W. Bush, Donald Rumsfeld or Richard Cheney. But what about the rest of us? How do average Americans, unaware of statistical techniques, view such "over the top" numbers as 654,965? Well, the only numbers they had heard before were numbers between roughly 30,000 and (rarely) 100,000 -- estimates drawn from very small samples of morgues and hospitals (and we've discussed the reasons why these are poor sources for estimates in previous blog postings, haven't we?) ... plus WAGs (wild ass guesses in military jargon).

General George Casey's comment is typical -- and we can all understand why and how he said it: "That 650,000 number [sic] seems way, way beyond any number that I have seen. I've not seen a number higher than 50,000. And so, I don't give it that much credibility at all." A normal statement by a normal person ... a normal person unfamiliar with scientific reasoning!

It is a truism that the sudden insertion of numbers that are beyond the range of even the outliers in previous estimates will always be taken with some skepticism by the general public, no matter how exact they are estimated or how "good" the methodologies used. My brother thought I was "pushing his buttons" ( habit of mine anyway) when I stated that the estimates used in the Johns Hopkins study (funded by MIT and published in the peer reviewed journal, The Lancet) were likely correct to within 50,000 if they randomly sampled 1,849 households throughout Iraq and were able to see 92% of the actual death certificates to eliminate vague guesses or double counting.

These numbers were estimated using the Normal Distribution for confidence interval and to be very specific, the researchers -- including eight Iraqi doctors and the university staffs -- are 95% certain that the actual number of deaths caused by Bush's War lay between 392,979 and 942,636 ... rounding to whole bodies.

To add to the credibility of the report, nearly 60% of the dead were men and boys between the ages of 15 and 44 -- not the norm in a population not at war.

Basically, the results are arrived at by comparing actual deaths between 2003 and 2006 with those between 2000 and 2003 ... and subtracting. That's a very simplified explanation but the full details of the statistical techniques employed can be found in The Lancet results. At the bottom of the article is a link to the PDF file of the relevant appendices ... for the more mathematically inclined among you.

[Just for the record, I picked up a second Master's Degree before going on for the Doctorate. That second MS was in "Experimental Inferential Statistics." The MIT and US Government financed Johns Hopkins Bloomberg School of Public Health study was done properly!]