« Terrorist enabler Lynne Stewart is sentenced | Home | Lynne Stewart Is a Traitor »
October 18, 2006
The Lancet Report
Just in case you want to know why the current Lancet Report is completely flawed in its methodology, read this. Steven Moore deconstructs the flawed statistical metholodogy employed by the researchers.
It comes down to two factors, far too few cluster points for its sample and no comparison of the statistical analysis from their report to a known demographic instrument, such as a census. And don't forget, that heirs to the British system, before the invasion, the Iraqis kept meticulous records.
In fact, they're indicted by their own appendix.
Bear in mind, they used 48 cluster points for a population of 27 million:
Appendix A of the Johns Hopkins survey, for example, cites several other studies of mortality in war zones, and uses the citations to validate the group's use of cluster sampling. One study is by the International Rescue Committee in the Democratic Republic of Congo, which used 750 cluster points. Harvard's School of Public Health, in a 1992 survey of Iraq, used 271 cluster points. Another study in Kosovo cites the use of 50 cluster points, but this was for a population of just 1.6 million, compared to Iraq's 27 million.
Alcibiades | 10/18/06 at 11:03 AM | Categories: - Iraq
Trackback Pings
TrackBack URL for this entry:
http://www.keshertalk.com/cgi-bin/mtb.cgi/5728
Comments
Moore's criticism is innumerate
Tim Lambert | October 18, 2006 02:05 PM
I do not care about cluster points. They are irrelevant to the faults of this survey.
Any survey depends on:
1) The use of random sampling. Any sampling bias, intentional or not, screws the whole thing up. Bias can slip in through the most innocent errors of the survey taker, for example, in a consumer survey being done in a mall, the appearance of the surveyor will impact the responses given.
2) Correct and Honest reporting by the sampled population. Bear this in mind, we have a population here that has no established history of responding to questions in that narrow frame of thought we westerners call "objective truth".
3) Correct and Honest reporting by the surveyor.
There are no controls for this, whatsoever. This renders the report worthless even if the there are no biases at all and every question was answered correctly and honestly. I can ask questions all day, nod my head dutifully with each answer, and still lie through my teeth about the results to back up my beliefs- if there is no outside party collecting the same information.
People give screwy answers to survey questions. I saw a corporate employee survey recently. It was in the form of the "do you agree or disagree with, answer 1-5 with 1 being disagree strongly" etc etc. The company posted the top 10 points of agreement and the bottom 10. The most agreed to point was that the leadership of their department was of high quality. The LEAST agreed point was "my department is run smoothly and efficiently". Forehead slap. Either an entire corporate workforce had absolutely no idea that the number one sign of a good department leader is a smooth and efficeiently run department, or an entire survey had been rendered utterly worthless, courtesy of point 2 above.
And not a cluster in sight. My prediction: If you exhaustively canvassed EVERY SINGLE IRAQI you wouldn't get anywhere near correct numbers.
Ben
Ben | October 18, 2006 04:24 PM
![[TypeKey Profile Page]](http://www.keshertalk.com/nav-commenters.gif)
