PJ Media

Elephant man (I think) asked about the Moore article in the WSJ. I responded to that but it was not published, hopefully because it was too technical.

Net result, regardless of the sampling frame which i would not accept from a 1st year grad student, is the lack of release of the full questionnaire, and the fault of not collecting age & sex (minimally) of the respondents.

Collection of serious “disaster” data requires, at a minimum:

Who lived here before X happened?
Who lives here now?
Where are the people before X and now?

These questions must be asked of the senior household head, not just “somebody” who happens to be available. That is why the age and sex of the informant (respondent) is so important.

In addition, the “cluster sampling” method has been severely abused by this charlatan. Cluster Samples (I’m not sure who originally designed them), are only indicated when one wants to look at “incidence” of some variable. Clustering is NOT a Time 1 – Time 2 system (which measures change on THE SAME VARIABLE over time).

The simple extrapolation of data from the Sadamm(Iraqi) census till today is simply beyond any demographic methodology, which takes into account changes in environment, policies and…as a matter of fact, cannot account for war.

Political? i can’t tell, and couldn’t care less.

From a methodological point of view the man is a dunce and charlatan. He also claims that the Appendices are the product of his “grad students”.

My opinions about Iraq are: it was a mistake, we (the US screwed up and gotta have a way out), BUT bullshit data have no role in policy decisions.

Sampling ANYTHING, during a war is a futile activity. Having worked in survey samples all over the world, I would NOT have accepted this contract. There are NO statistical parameters upon which to base change.

Better this (no – name dipper from Hopkins) pay attention to the basics of methodology).

Better rely on the Ernie Pyle’s.