Monday, October 16, 2006

Innumerate vs. Non-innumerate criticisms of the John Hopkins Mortality Study

There's been some confusion about this in the comment threads. So let me set this out:

*Criticism that the sampling exhibited various biases. (NOT INNUMERATE)

*Asking where all the bodies are. (INNUMERATE).

*Making reference to the wide confidence interval. (NOT INNUMERATE)

*Referring to Iraq Body Count or other media-report-based sample as disproof (INNUMERATE)

*Questioning whether the Sampling Protocols could have been executed flawlessly in the circumstances of Iraq (NOT INNUMERATE).

*Refusing to believe the number because it just seems too high or because there is a midterm election in the US next month, or because the Iraqi and American governments don't like it or because snooty leftists have referred to it. (INNUMERATE)

You're welcome.

Pithlord approved links:

Andrew Gelman

Tim Lambert (any of last ten posts)

The BBC (includes Iraq Body Count criticsms and responses by study authors to other criticisms)

Iraq Body Count's own press release

Mike Dunford, Take 1 (critical of the John Hopkins study) and Mike Dunford, Take 2 (reconsidering some of his criticisms

Dunford Take 2 walks back from his original criticism that the sample was biased to urban areas, although he doesn't say why.

No comments: