Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Politics for Pros- moderated -- Ignore unavailable to you. Want to Upgrade?

To: Alastair McIntosh who wrote (154799)	1/15/2006 11:49:36 AM
From: TimF	Respond to of 793718

I looked at d-n-i.net It says things like "visited 33 "clusters" selected on an entirely random basis" and "again selected by rigorously random means", but it doesn't really address the concerns of those who have called them insufficiently random, it just asserts they are perfect. It also revisits the type of comments you already posted. At least it gives direct quotes like - "I have the highest respect for the rigor of the sampling method used and the meticulous and courageous collection of the data." I'm not big on arguments from authority esp. in controversial, political charged discussions. As a statistician we can assume that Pierre Sprey should know what he is talking about, but we can't assume that he is unbiased (the fact that he is "CounterPunch's consultant statistician" would seem to make bias likely) or that he is correct. I would listen to the arguments of someone who would seem to be a legitimate authority in a relevant area, but I would not just accept his or her conclusions automatically. The only actual argument he makes is about Gaussian distribution and how it is not necessary. It might have some relevance to the argument but only a minor one. It doesn't effectively respond to the arguments made about the pre-war Iraqi death rate, possible non-random sample, the fact that only a small part of the data from the sample was checked by asking for death certificates or other forms of evidence, or the fact that their isn't any evidence other than this study of that type of death increase. Consider that the study is not even talking about deaths to today but rather to September 2004. No other source even remotely hints at that type of death total per day. In fact there isn't any indication that the number of average people who were supposedly killed per day by Pierre Sprey's upwardly revised estimate, or even the original studies 100,000 estimate were killed on ANY day if you are only counting civilians (and the study claims 100,000 civilians were killed). Turning to timlambert.org " Unfortunately, Fumento seems to have missed the immediately preceding sentences in the Lancet paper, where they noted that, when asked, 81% confirmed with death certificates" Yes 81% of the households that were asked. But only 73 households were asked and only 63 of those provided the confirmation. A thin reed on which to base an extrapolation to all of Iraq. " That’s right, they properly excluded the outlier Falluja in their estimate of 98,000 and Fumento didn’t notice this fact." Which might be an effective put down of Fumento's specific argument but which doesn't settle the larger issues with the sample. Timlambert's site also explains how Iraq's death rate could possibly be lower than that of the US (the country is younger). but it doesn't provide much to support that specific death rate as accurate. And apparently the claimed death rate was higher pre-2002 than in 2002. I find it somewhat questionable that there was such a large improvement just before our invasion, and even if the data was accurate its questionable that the improvement would be sustained and not just a fluctuation over a relatively short time period.

To: Alastair McIntosh who wrote (154799)	1/15/2006 12:02:06 PM
From: TimF	Respond to of 793718

The three biggest issues are IMO the accuracy of the claims for the prewar death rate, the lack of any other evidence suggesting an increase in the death rate even remotely approaching that argued for by the Lancet study, and the issue of sampleing. I think the third issue is explained pretty well at seixon.com or quickly and directly by seixon.com Note the areas that where not sampled. I didn't see anything in either of your links that seemed to even try to deal with this issue. ---------------------------- I did see a link to sagenz.typepad.com followed by an effective response to the original argument made at that link (which was an argument that I hadn't brought up), but the blogger corrected the article and it has some interesting data. Apparently the pre-war death rate estimate was from the sample but it relied on the deaths of just 12 children, just 4 non-elderly adult men to project out the death rate for all of Iraq for those groups. I understand that a random sample can be relatively small and still provide useful information about a larger hole but I doubt very much that a sample that small is sufficent. Tim