Lancet-go-round

Tim Lambert easily dispenses with criticism of the Hopkins/Lancet study of "excess deaths" in Iraq grounded in the divergence between their estimate of Iraq's pre-war death rate and the UN's estimate. The UN's estimate turns out to have just been a kind of guess with no real methodological grounding at all. Lambert's countercriticism of the "main street bias" line of attack doesn't seem quite as airtight to me (it's a little bit unclear based on the description exactly how this survey was done), but that was always a fairly speculative criticism (in that nobody really knew whether or not such a bias even existed, and other elements of the survey were biasing the death count in the other direction) so what the Hopkins team did still seems like by far the best estimate available. Better, certainly, than the "these numbers are very high so I choose to ignore them" method preferred by many.

Obviously, if the American and British governments -- or conservative think tanks and media outlets -- genuinely feel that the Hopkins team's methods were unsound, there's an obvious solution available to them: Design a method for a different comprehensive study of Iraqi mortality and fund its implementation. This is a sufficiently important question, and sufficiently difficult to pin down precisely, that it would make perfect sense for several different studies to be conducted.