StatsGuy writes about the Lancet study (via Kathy). I'm just going to include a small snippet, this a long post with lots of stats and survey stuff that you should really read through:
The interview team went to 1849 households in urban areas of Iraq and encountered only 15 refusals and only 16 residences where neither the head of the household nor a spouse was in. Don’t forget that they only went to each household once: there was no follow-up whatever. If I ran a door-to-door survey with a response rate of 98.3% on the first go-round, I’d think I’d died and gone to statisticians’ heaven. That is nothing short of miraculous. That response rate implies that family heads in urban Iraq are virtually always at home.
Reading through the following paragraphs should really capture you interest -- the survey team was probably one of the most speediest and effective teams ever fielded by humanity. Or perhaps something else was going on.
Stats guy also writes (in a separate post) that the Lancet report authors ignored other studies on Iraq:
This I found odd. Articles in academic and professional journals that address topics of controversy generally include references to previously published studies and discuss the perspective the current article takes vis-à-vis the views and findings of those earlier studies. That is how scientific knowledge advances—by critically engaging published findings of other scholars and specialists.
The authors of the 2006 Lancet article, however, appear uninterested in critical engagement with the ILCS estimate of Iraqi deaths. Yet we know that the Lancet researchers are aware of the ILCS, for they refer to it twice in their footnotes. The first page mentions "surveys that assessed the burden of conflict on the population" and the fact that "insufficient water supplies, non-functional sewerage, and restricted electricity supply . . . create health hazards", and for these the ILCS is footnoted.
But as for critical discussion of the enormous difference between the ILCS estimate of deaths and the estimates generated from both Lancet surveys, the authors don't want to touch that. They don't even acknowledge its existence.