Jousting with the Lancet: Pajamas Media Interviews Professor Gilbert Burnham
There is no political bias, but there could be statistical bias (oversampling in Baghdad would likely reduce the size of the estimates as it is not a high-mortality geography). If you read the Lancet, which is not exactly a political journal, you would see that it also reports a confidence interval, which means that we can be 95% sure that the actual number of deaths in post-invasion Iraq is between 392,979 – 942,636. This wide, as is openly stated in the article, but makes all of this hooplah about 654,965 seem ridiculous. You may not like the results, but if you don’t believe them then you should refuse to accept most published polls. You must therefore publically take the following positions: The Democrats are not going to make gains in congress, the president is not less popular, the average price of gas is not going down (yes, that is a geographically clustered sample), and there is no reliable estimate of the US native born population. You can complain about the study but disregarding it is willful ignorance.
