And if you've ever seen the Faculty Lounge here at ATS-Orlando, you know where the money goes. Have you seen Brian Russell's company car?
Anyway, The Lancet conducted a study (registration required) of Iraqi deaths and concluded that approximately 654,965 have died since the beginning of the war (h/t). Take your time and read the sampling and data collection methods carefully.
The researchers randomly selected 50 population clusters and then later eliminated data from three due to a mapping error. Within these clusters, researchers approached 1,847 homes and asked residents about known deaths. There were 545 death reports, of which 501 death certificates were present. From the former number, the researchers extrapolated 654,965 deaths across Iraq.
As Sexion (h/t) points out, if death certificates are the count of deaths in Iraq, then why not just ask the Iraqi government how many death certificates were printed during a certain period of time? This would be a more reliable form of data collection and would form a more representative (if not approaching total) sample.
Why not? Because the Iraqi government says that The Lancet's claim is nonsense. The existence of the official copies of the death certificates can be visually verified. Using The Lancet's method, we can rely only on the word of the researchers.
And where are the bodies? Iraq has a population of 27 million, so this study asserts the death of 2.5% of the population. That's a lot of corpses. Any idea where they went?
On a side note, it's rather odd that The Lancet only releases these studies on the eve of major American elections. Coincidence? I'm sure that Lancet editor Dr. Richard Horton, who recently spoke at a pro-Hezbollah rally, will deny any political motives.