To Tom Blumer: that’s the problem with you anti-science, anti-research types — you can’t recognize the difference between stuff that came from serious research and stuff that came from someone’s butt, and that National Journal piece is crap on a stick. The fact is that you’re comparing the results of a long, careful John Hopkins investigation from just this past February to a an article appearing in a second rate conservative magazine (its nonpartisan claim is BS) not exactly known to break stories.
A key admission in that National Journal piece goes: “With Iraq’s central government barely functioning, health services overwhelmed, and political agendas coloring all agencies, no reliable statistics exist so far.
“The Lancet study was based on techniques developed by public health experts to determine rates of illness and death from epidemics and famines in large populations. This ‘cluster’ sampling is a relatively new methodology that attempts to replicate the logic of public opinion polling in Third World locales that lack a telecommunications infrastructure.”
If you think this through, then you should realize that this is essentially saying that the methodology used by Iraq Body Count — primarily news sources — and which Bush had quietly used, would produce the most severe undercount because few reporters are going to venture into the really bad areas of a war zone just to count bodies. But that’s where most of the deaths are going to be, so…. That National Journal piece, however, blithely ignores this in favor of plunging ahead with a lot of dopey reasoning not based on anything other than ignorance of how those types of surveys are conducted in places like Iraq, and how the data collected is processed to account for all possible major variables.
A second survey by a UK polling firm called ORB from January of last year raised the civilian casualties even further to about a million, which is about what you would expect if you extrapolate from the 2006 Lancet results and my follow-up statistical analysis.
If you have any friends with math skills, you can have them dupe my statistical results easily enough (my first was done during a Usenet argument): Take the population of Iraq, stick in any murder rate you want from something really low like 1/100,000 (Saudi Arabia, Oman) to a high of say 70/100,000 (Columbia) as an estimate for the prewar violent death rate in Iraq as the baseline. Then from March 20, 2003 to now, try factoring in one of the lower death tallies, like the IBC one, and see how much that raised the average violent death rate since the invasion and compare that with crime statistics of any American city over the past couple of decades and see if the numbers make any sense at all. After doing that, try plugging in a count of 1,000,000 casualties and see if the results then make a little bit more sense for a country at war and not just a place where you just need to be little bit cautious when walking alone at night.
