Donna Ferentes said:
Useful
piece explaining why objections are largely spurious.
I read that article earlier but maybe you could remind me what it says about the following problems with the figures:
1. The mis-match between the lower "baseline" (pre-invasion) figures and higher UN, NGO and anti-embargo-campainger figures about mortaility rates prior to the invasion.
2. Problems with actually being able to do proper random sampling when the census figures are so unreliable, when populations have been moving around and when the security situation is so hostile.
3. In interpreting the figures - the stark discontinuities across regions of Iraq ie far higher to the north and west of Baghdad, by age, gender and other group - ends up giving utterly ridiculous figures for the worst affected areas. Obviously something has gone wrong in this study.
4. Implicit assumptions of bell curves and normal distributions contained within the study.
5. The lack of bodies: where are they?
6. Given that the study claims that 92% of deaths were backed up by death certificates, why is there such a large mismatch between these numbers and the Iraqi ministry of health and figures from other Iraqi bodies - the very bodies that issue death certiicates in the first place. There is something going here that doesn't add up, yet the study doesn't even address this at all, simply taking certificates at face value.
7. The total lack of analysis of possible motivations for over reporting of fatalities - for example political views, an awareness of the reasons behind the survey and possible compensation for fatalities. The survey simply dismisses a whole range of questions out of hand, assuming that pe3ople are giving accurate information or are being totally honest.
8. Finally, and most importantly, the study itself hints at various conclusions being drawn - about cause and effect, the role of US/UK forces and various other value judgements and political viewpoints whicn are largely out of place within a so-called neutral survey.
This is taken even further by the media and people with a view on the desirability or otherwise of US/UK foreign policy - in other words the survey itself seems to contain bias and a 'view' in its commentary and jumps to various conclusions within its text. Various people have already started using this survey to make claims that go beyond what even the survey itself makes, and the Lancet's editor is looking more and more like a tabloid editor rather than impartial medical science publisher.
The problems with this survey don't primarily revolve around details regression analysis or the statistical techniques used, and you don't have to look very far in science to see people making massive mistakes because they simply plough ahead with theoretically "correct" techniques but mess up some other very basic parametres and reality checks. Pseudo-scientists often are the most dangerous as they are so blind to the wider context and such unquestioning faith in their mechanical 'techniques' that they produce utter drivel without even realising or admitting they have done so.
just doesn't quite do it for me any more.
Nothing to do with the methodology.
