Slate magazine's Fred Kaplan takes a hard look at the statistics behind the Lancet study on Iraqi deaths due to the war.  The study's figure of 655,000 has proved controversial.  Kaplan lays out a case why we should treat the number with some skepticism. He also makes the following point which is, I believe, well taken:

Here lies the danger of studies that overstate a war's death toll. Thewar's supporters and apologists latch on to the inevitable debunkingsand proclaim that really "only 100,000" or "only 200,000" people havedied. It's obsceneit sullies and coarsens the political culturetoplace the word "only" in front of such numbers.