Please visit the new home of Majikthise at

« Tim Blair and the web of belief | Main | Help Biting Beaver: A call to arms »

October 14, 2006

Brad DeLong on the Lancet study

Brad DeLong takes on William Arkin of the Washington Post over the Lancet study of Iraqi deaths since the US invasion.

Arkin says that the Lancet authors used an unrealistically low baseline mortality rate. The Lancet authors estimated excess deaths, i.e., deaths that would not have occurred if the baseline mortality rate had stayed the same. So, choosing an artificially low baseline would widen the gap between pre- and post-invasion death rates and therefore inflate the estimate of excess deaths.

In fact, as DeLong points out, Iraq had a very low crude death rate before the because its population is very young. The baseline death rate the Lancet authors assumed for Iraq is right in line with the current crude death rates of neighboring countries.

Meanwhile, Daniel Davies continues to wage war on innumerate hackery on the Guardian's blog.


TrackBack URL for this entry:

Listed below are links to weblogs that reference Brad DeLong on the Lancet study:


Numbers like this are just too important to be dismissed. This recalls the arguments of holocaust deniers. It’s more important (to me) that there is a true number and how we arrive at that number. Until the disparity between the Lancett numbers and the numbers accepted by both right & left think tanks are somehow reconciled I am suspending judgment as to the truth of things.
When a study ends up with ten fold the excepted numbers, one has to question that studies veracity. The truth is more important than partisan/ideological weaponry. There are ways to confirm the actual number. 600,000 plus requires mass graves, funerals, families and loved ones that can produce photographs, birth certificates, old I.D. ect. For this gigantic number to float around without further investigation and critique is a crime against humanity. We are all aware of the arguments over Pol Pots death toll or Moe’s Cultural Revolution. These kind of numbers being argued is dangerous in and of itself. They become assaults on truth (the obtain-ability of truth) on a subject that goes to the core of our humanity. One hopes this sort of thing does not become an endless historical argument of ideologists.

Be the numbers greater or smaller (or in between) The truth of it should be paramount.

Well, to be precise, it's not just that Iraq has a young population. It's also that its level of development is far higher than this of the average third-world country. In Saudi Arabia and Kuwait, which didn't have to cope with sanctions and a crippling occupation, the death rate is even lower. The only Arab Middle Eastern country with a higher death rate than pre-war Iraq, Yemen, is also the least developed in the region (in fact, Haiti narrowly beats it for least developed non-African country).

Good Math Bad Math, one of the Seed blogs, takes the hot air out of the bizarre "logic" of several of the poor conservative commenters who have to make it all add up to "no we didn't just murder half a million people"

Moe's cultural revolution. That is genius-level work, man. Genius-level. Buy yourself a beer.

Arkin was wrong: the second Lancet study derived its own pre-war baseline from its interviews. It only used standard figures (ie, CIA) to validate their own. They took comfort that their own figure (5.5/1000) was comparable. Arkin clearly didn't read the study, and you're repeating his error here re the excess death calculation.

Yes Moe…as in Larry Curly &.

The stooges led western civilization to turn itself upside down & inside out in the name of radical egalitarianism.

Instinctively you give to me
The love that I need
I cherish the moments with you
Respectfully I say to thee
I'm aware that you're cheating
When no one makes me feel like you do

Phitz, when you channel Diana Ross, it makes me feel all squishy inside.

This criticism of the Lancet study is from the folks at, who strongly oppose the war, so they've got no partisan axe to grind on this one. They offer some "reality check"-style reasons why they find the Lancet estimate to be way, way too high.

Alon, the link you provide argues that another survey (the ILCS) is invalid. It fails to address even a single one of the several criticisms of the Lancet study that IBC makes here.

Ah, that criticism is just a big argument from incredulity. The media didn't initially report that many civilian dead in Vietnam, either. I was referring to a more serious IBC criticism, which brings a legitimate study.

The "more people are killed by the occupation than were by the war" part is a bit of a surprise, but not that much. Even the reported figures show a steady increase in the number of dead; the number of people who died today in Iraq may be lower than the number who died on 2003/3/21, but Shock and Awe lasted a little more than a month, while the current violence is continuous. If I'm not mistaken, the IBC count when Bush did the mission accomplished photoop was still in the four figures, so even it would admit that the occupation has killed more civilians than the war.

I have a discussion of the IBC "reality check" here. Alon Levy is right; it's an argument from incredulity, with a side order of "that would be surprising if it happened in a modern, industrialized nation *not* on the brink of civil war".

The comments to this entry are closed.