UPDATE: Sorry I got your name wrong, David.
ADDED LATER: This article by Nick Cohen gives one reason for my prediction. It's actually advocating the replacement of Tony Blair with Gordon Brown, but my eye was caught by this paragraph:
For reasons no one can adequately explain, the opinion polls have underestimated Tory support for years. Notoriously in 1992 they failed to predict Major's victory. What isn't as well known is that in 1997 the average opinion poll result for the Tories over the campaign was 3.5 per cent below their actual vote. In 2001 the polls underestimated their real support by 6.5 per cent.I can explain it. What's more I could explain it in 1992 before the election. Alas, I lacked the intiative to actually go into a betting shop. (To this day I never have entered one of these houses of sin.) The explanation is that the media, talking heads, academics, comedians and the like, portray the Conservatives, or Republicans in the US, as the party of those who are unfeeling and/or uneducated. That makes some Conservatives or Republicans a little more reluctant to tell pollsters how they plan to vote.
(Hat tip to James Rummel of Hell in a Handbasket.)
Shannon Love's posts and the comments to them engage with this post by Daniel of Crooked Timber and its comments. Later on Chris Bertram, also of Crooked Timber, posted this follow up about the morality of air strikes, also attended by comments from Shannon Love and others.
ANNOYED BY COMPUTER STUPIDITY UPDATE: I tried to sumbit a comment to Ginny's Chicago Boyz post but it was rejected by the system due to "questionable content." I am mystified. Perhaps the program automatically rejects various WWII keywords as likely to degrade debate. Anyway, I'll assuage my frustration by posting my comment here:
An anonymous commenter says (in reply to John F): "the real issue is the capacity of modern warfare bombs, which is many, many, maby times the capacity of WWII bombs."
Well, no. While it's of course true that we can now build bigger bombs than those used at Hiroshima and Nagasaki, they haven't been used.
The defining difference between the bombs actually used in Iraq and WWII bombs is that modern technology can deliver bombs/missiles much more accurately.
So, as John F says, one would expect fewer unintended victims.

From what I know of the two I am much more inclined towards the Worstall view of the world than the Lambert one. But in the interest of a detailed and highly educated debate about statistics that will almost certainly not include me as a participant, there they both are.
UPDATE: More from Ragout, concentrating on Beth Osborne Daponte, the demographer Fred Kaplan of Slate spoke to.
ANOTHER UPDATE: Tim Worstall says he was wrong and Tim Lambert was right. I still love you, Tim! Thank you, Other Tim!
Time for my totally unscientific bit. I stand by my prediction that the study will be withdrawn or corrected. I don't know much about statistics* but I know something about history, military history in particular. A hundred thousand is just way too high. That's big war stuff and this is a little war.
I do not, of course, expect anyone on this earth to believe the Lancet study must be wrong because Natalie's gut feeling says so. (Or because Natalie's husband, the War Studies Guy, says so too.) For all that, my argument is not as irrational as it first appears. There are many fields - ironically, medicine among them - where even world experts intuit first and back it up afterwards. A "nose", antique dealers call it.
*Here's an odd thought, not intended to be contentious. I have A-Levels in Mathematics and Further Mathematics and also did some statistics as part of a degree in Physics. Hence I probably do know more statistics than 99% of the population. But I still say "I don't know much about statistics" and mean it. Complicated world out there.
Legislation in Oklahoma which allowed the home-owner to use force no matter how slight the threat has reduced burglary by nearly half since it was passed 15 years ago. What British police condemn as "vigilante" behaviour has produced an American burglary rate less than half the English rate. And, while 53 per cent of English burglaries occur when someone is at home, only 13 per cent do in America, where burglars admit to fearing armed home-owners more than the police. Violent crime in the US is at a 30-year low.Whatever became of the Englishman's castle? He did not lose the right and means to protect himself at once. It was teased away over the course of some 80 years by governments claiming to be fighting crime, but actually fearful of revolution and disorder. When the policy began, crime was rare. For almost 500 years, until 1954, England and Wales enjoyed a declining rate of violent crime. In the last years of the 19th century, when there were no restrictions on guns, there was just one handgun homicide a year in a population of 30 million people. In 1904 there were only four armed robberies in London, then the largest city in the world.