Semi-Daily Journal Archive

The Blogspot archive of the weblog of J. Bradford DeLong, Professor of Economics and Chair of the PEIS major at U.C. Berkeley, a Research Associate of the National Bureau of Economic Research, and former Deputy Assistant Secretary of the U.S. Treasury.

Sunday, October 22, 2006

Why Oh Why Can't We Have a Better Press Corps? (Yet Another Wall Street Journal Edition)

Yet more lies from the editorial page of the Wall Street Journal:

OpinionJournal - Featured Article: 655,000 War Dead? A bogus study on Iraq casualties. BY STEVEN E. MOORE: I was surprised to read that a study by a group from Johns Hopkins University claims that 655,000 Iraqis have died as a result of the war.... [T]he key to the validity of cluster sampling is to use enough cluster points.... What happens when you don't use enough cluster points in a survey? You get crazy results....

That's a lie: you don't get "crazy" results--you get imprecise results with large reported standard errors. The 95% confidence interval for Johns Hopkins/Lancet is 500,000 deaths wide. It would have been only 170,000 wide if they had had 500 rather than 50 clustesr.

There was a perfect example of this two years ago. The UNDP's survey, in April and May 2004, estimated between 18,000 and 29,000 Iraqi civilian deaths due to the war. This survey was conducted four months prior to another, earlier study [Lancet I] by the Johns Hopkins team, which used 33 cluster points and estimated between 69,000 and 155,000 civilian deaths--four to five times as high...

Tim Lambert calculated that the UNDP study which (a) covered only 2/3 as long a period, and (b) covers a much smaller class of deaths is in no wise inconsistent with Lancet I:

Iraq Mortality | Iraq Mortality: Using figures given in the Lancet Study, Australian academic Tim Lambert derives a figure of 33,000 excess Iraqi deaths due to military action by insurgents and US-led forces in the period up to September 2004.... By simple multiplication, this results in 33,000 "war-related" deaths according to the UNDP definition of this term, for the period March 2003-September 2004. This compares to the UNDP estimate for narrowly-war-related violent deaths of 24,000 for the period March 2003 to April 2004.... If we crudely scale up the UNDP figure to take account of the longer Lancet time period, we reach a figure (33,000) which is exactly the Lancet-derived figure of 33,000 violent deaths due to military action. So, rather than there being a conflict between the two survey-based estimates, as often suggested by officials and apologists for the war, we see mutual confirmation....

Steven E. Moore in the Wall Street Journal continues to blather:

Curious about the kind of people who would have the chutzpah to claim to a national audience that this kind of research was methodologically sound.... 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...

As everybody literate in statistics knows--and as Steven Moore doesn't--it's not the number of clusters in proportion to the that is important, it is the absolute number of clusters. Steven Moore: Ignoramus who doesn't know basic statistics? Liar who bets his readers don't know it is the absolute number of clusters is important? I don't know.

Why oh why can't we have a better press corps?


Post a Comment

Links to this post:

Create a Link

<< Home