« The Cultural Cold War | Main | China and Iran »

October 29, 2004

Comments

James Hamilton

I'm going to read this report with great care before committing myself: the Lancet deserves that much. I am concerned on three counts. One, the accelerated peer review; two, the question of peer review when new methodology is used, as in this case; 3, the sheer size of the figure when related to the scale of military activity - 100,000 dead is not far away from a Hamburg or Dresden, and I need to be convinced that this is what we've had. A sample of 998 families is large enough to draw reasonable data, but I want to see how they've factored in the "extended family" factor. On the positive side, they do say that they have found relatively little sign of wrongdoing by US troops on the ground - were this a politicised research paper, I wonder whether they'd have bothered to make that clear?

Eamonn

It is simply ridiculous to say that American airstrikes have killed 50,000 or more Iraqi civilians. If this were true, Bush and Blair would be thrown out and the coalition would leave Iraq with their tails between their legs. I think the methodology of the study will require a "little" analysis. Remember the Lancet also published the unfounded claims by Andrew Wakefield about MMR and autism. Shame on the Lancet.

But of course in the political sphere the left-liberal Lancet does want to keep up with the left liberal BMJ:-

http://bmj.bmjjournals.com/cgi/content/full/329/7471/924


James Hamilton

I'd still say, read the report first. Which I haven't, yet. The TCS article makes a valid point - the peer reviewing of this new paper has been fast-tracked to an altogether extraordinary degree - and that the reasons for this are probably political. But that on its own isn't enough to derail what the paper actually claims, not by a long way. For that, we'd need clear misuse of statistical methods, incorrect collection or analysis of data, etc., and if those are lacking, some of us (including myself) will be facing unpleasant facts that we'd rather weren't true.

John F

I don't believe this for a moment.
IIRC the total civilian casualties for Italy WW2 were around 70,000.
And the timing...
If this isn't a political ploy I'll eat a copy.

Anthony

I'm reading the full paper now. Although I am no master statistician, there are issues about using cluster sampling as a method - in terms of precision - as well as the fact that in the results section of the paper they state that "We estimate that there were 98000 extra deaths (95% CI 8000 - 194000) during the postwar period ..."

This confidence interval is wide enough to cover the lower Iraqi Body count figure - which has been criticised widely, but is at least based on a verifiable body count, rather than cluster sampling.

The 95% CI is the range of values within which we can be 95% sure that the true value lies for the whole population from whom the study population was selected. So these authors are 95% sure that the number of casualties was between eight thousand and one-hundred-and-ninety-eight thousand. This nuance is lost in the news reports which state that "Poor planning, air strikes by coalition forces and a "climate of violence" have led to more than 100,000 extra deaths in Iraq, scientists say."

What the scientists have is a measure of the uncertainty of the figure (between 8000 and 198,000), they do not have a measure of the certainty of the figure actually being 100,000.

Eamonn,

I made a couple of responses to that awful BMJ piece, and got lumped into the Zionist Lobby...

http://bmj.bmjjournals.com/cgi/eletters/329/7471/924#79896

http://bmj.bmjjournals.com/cgi/eletters/329/7471/924#80325

I have yet to work out how the Zionist Lobby controls my actions, but I believe I heard one Malaysian leader suggest it was something to do with the Jewish inventions of democracy and human rights which they use to trick others into to thinking they should be treated as others are...

The comments to this entry are closed.