I’ve not yet seen the study, as it is due to hit the Lancet website on Thursday. I’m going to be all over it, examining its methodology and discussing it in more depth. I am very familiar with this group’s first study, which estimated (at the low end) that about 98,000 Iraqis died in about the first year post-invasion. Why a low estimate? Because they came to that number by excluding Falluja from the analysis. They figured that so many people had died in Falluja that including it would bias their sample. Including Falluja would have brought their figures to about 200,000 dead.
The researchers reportedly used cluster sampling again in the new study, which is the same method used in the prior study. It is the appropriate method when investigating such topics as this. In the first (and I’m guessing also in the latest) study, they randomly selected a number of clusters/areas to visit, then were randomly assigned to neighborhoods in these areas through random numbers entered into their GPS systems. They went to the 30 nearest houses and knocked on doors, collecting data regarding deaths prior to the invasion (starting 1/1/2002 until the start of the invasion) and post-invasion. They requested death certificates (to confirm the report of death) from a relatively small number of families, and these certificates were provided over 80% of the time. In the new study, they were reportedly provided over 90% of the time when requested. You may ask why they did not request death certificates from everyone – consider that it may be considered disrespectful and could result in danger to the researchers to ask for confirmation of household deaths.
A test of the American media is coming. Will it dare report these figures? If it does, can they resist the temptation to include accompanying commentary from a Republican blowhard who moans about how this study is "just politics," is "biased," and just an attempt by liberals to "embolden the enemy"? When the war was relatively popular in 2004, the media did not linger long on Roberts et al.’s findings, preferring to refer to other sources (such as www.iraqbodycount.net) for its reports of total Iraqi dead. Would the American people support a war in which an estimated 650,000 Iraqis have lost their lives? Granted, many of the deaths are due to insurgent and sectarian violence. But remember that such groups were not killing people prior to the invasion. Saddam was one bad apple, but even he likely could not have dreamed of replicating this type of chaos and killing.
Update: The new study just posted on the Lancet's site. More upcoming, hopefully soon...