That Lancet Study! The one about 100,000 dead Iraqis
The Age of Unreason
The Blogosphere is abuzz today about the Lancet study claiming that 100,000 Iraqis have been killed since the beginning of Operation Iraqi Freedom.
instapundit has a long and growing post on the subject. This blogstorm was initiated by this Slate article debunking the original study.
slate.msn.com
I have three comments to make about this.
Firstly the use of a scientific publication to make POLITICAL POINTS is reprehensible. This Lancet study was published just before the American Presidential Election and was clearly an attempt to make President Bush look bad just before he faced the electorate. Scientific Journals should be vehicles for publishing research and discussions on the nature of the natural world we inhabit and not ever be used to push a political (or any other ) agenda.
Agenda driven science threatens to undermine the scientific process that has proved so useful in advancing human knowledge and welfare over the past two centuries or so.
Secondly it has been noted that this study was based upon a 'Random Cluster' design and according to the Slate article.
<<<
One of the 33 clusters they selected happened to be in Fallujah, one of the most heavily bombed and shelled cities in all Iraq. Was it legitimate to extrapolate from a sample that included such an extreme case? More awkward yet, it turned out, two-thirds of all the violent deaths that the team recorded took place in the Fallujah cluster. They settled the dilemma by issuing two sets of figures—one with Fallujah, the other without. The estimate of 98,000 deaths is the extrapolation from the set that does not include Fallujah. What's the extrapolation for the set that does include Fallujah? They don't exactly say. Fallujah was nearly unique; it's impossible to figure out how to extrapolate from it. A question does arise, though: Is this difficulty a result of some peculiarity about the fighting in Fallujah? Or is it a result of some peculiarity in the survey's methodology?
>>>
Now I haven't read the Lancet article, something I fully intend to do, but if they didn't know how to handle the Fulluja cluster, as stated in the Slate article then they clearly have NO EXPERTISE in Sample Design. There is well established methodology for designing, selecting participants and analyzing the data collected data from Cluster Samples.
I don't want to get technical but under this methodology Fulluja should have been included!
The original Sample Design should have ensured that it was selected because it was known to be atypical! The original sample design would have accounted for the fact that the Falluja cluster was an atypical cluster and the analysis of the results would reflect this. There are sound and well established mathematical techniques to accomplish this.
Thirdly a bane of all sample surveys is the item called in the literature "Non Sampling error". One major source of non sampling error in surveys which collect responses from people is that people do not always tell the truth.
Famously surveys on how many cigarettes people smoke daily show far fewer cigarettes smoked than are known to be sold. Survey respondents typically understate the number of cigarettes they smoke daily, they are embarrassed to admit the truth, they may even be lying to themselves.
As I understand the Lancet survey, interviewers went to selected households and asked about family deaths in the period between the invasion of Iraq and the day of the interview. To say none (as most people probably did) is kind of boring so some respondents would be tempted to invent a dead relative. If the interviewee had an anti American agenda, as some surely did the temptation to lie would be even greater.
All in all I don't think this famous Lancet Paper is worth the paper it is written on but I shall definitely read it and if I change my mind I will be sure to tell all.
UPDATE:
Welcome to all Instapundit Readers.
I have now obtained the Original paper and will work my way through it. This is not an armchair exercise but is something that requires a desk, paper, pencils, calculator and reference books not to mention possible trips to the library. If the Slate article is way off beam I will be sure to let you all know.
posted by Andrei
sacredcowgraveyard.blogspot.com |
