It appears there is ongoing debate about the statistical approaches and methodology of the Lancet Iraq mortality study from 2004:
'Mortality before and after the 2003 invasion of Iraq: cluster sample survey' (from The Lancet, Volume 364, Issue 9448, Pages 1857-1864 by L. Roberts, R. Lafta, R. Garfield, J. Khudhairi, G. Burnham)
I can see why it would be politically expedient to shout it down: the first study reckoned around 100,000 Iraqi civilians had died as a result of the war, and its follow-ups seem to have revised that figure to getting on for 700,000.
So, if the study is sound, we can say with confidence that the best part of a million civilians have died as a direct or indirect result of the war in Iraq. This is ten times the estimate of the Iraq government (which broadly agreeds with the
Iraq Body Count total, which has to be treated as a minimum as it is based on confirmed deaths reported by the media). The higher number is also 70 times greater than the average American's perception of civilian deaths (according to a survey I could dig up a reference to if you wanted).
However, I simply don't have enough understanding of the issues involved to see if any of the objections raised have a genuine scientific basis.
There have been several objections. David Kane's paper seems to come up repeatedly, which is reproduced on this nice lady's blog
here in it's entirety.
It is somewhat amusing to read those who clearly haven't got the first idea about Kane's argument but who nonetheless support his approach enthusiastically. Similarly, there seem to be many commentators on the internet who find his approach fundamentally flawed, but I wonder what proportion of them truly know what he is talking about.
In the simplest of laymanish terms then, these seem to be the main clusters of objection to the study. I have no idea which have merit and which do not.
1) The cluster of data from Falluja is in some way problematic and shoud be excluded (or not) and somehow makes the overall number more than it should be (or not) - or possibly less reliable either way (can't see how this could happen)
2) That not enough areas were sampled for, or that there was a bias towards areas near major thoroughfares (which would have more fatalities?) thus hampering the search for a decent statistical result.
3) That the data isn't being made available by the researchers for checking by others.
(ABE: I have realised that I need to start referencing some of these - which I will do, if only as a memory aid to myself. Also, I will keep adding to this list here if I find more reasons and reserve the right to tart up my writing as I see fit to make my point clearer)
4) That the prewar mortality rate of 5.5 per 1000 per year was pulled out of someone's arse - and that according to the CIA factbook it was 6.5 or thereabouts, thus totally undermining the assumptions of the study.
I won't pretend to understand what's going on here, and there are so many shrill claims and counterclaims that I just don't know. Ideally I want someone to go systematically through the objections and tell me whether they are justified or not.
It is interesting to note that the British government actually thinks the study is
kosher. Are they wrong?
I'd also like to talk about the wider significance of this study.
(I had a search for relevant threads, but there didn't seem to be any.)
Edited by Tusko, : No reason given.
Edited by Tusko, : Another hyperlink
Edited by Tusko, : No reason given.