The Lancet -- Not the Sharpest Tool in the Shed

A British medical journal, The Lancet, has just "fast-tracked" to publication an article estimating that the Iraq war and its aftermath have seen about an additional 98,000 deaths in the past 18 months over what would have been expected based on official pre-War Iraqi statistics.

Uh, exqueaze me? Official Saddam Hussein Reign of Terror Totalitarian Dictatorship statistics were the mortality benchmark? Now, correct me if I'm wrong here, but I don't recall ANY of that madman's mass graves having little body-count booklets buried in them next to the 6-year-old girls with their heads shot off--still clutching their mothers' hands.

The authors' estimation procedure also leaves a lot to be desired. Remember, their number is an estimate, not an actual count. As Tim Worstall has pointed out at Tech Central Station, the authors guess that somewhere between 8,000 more people and 194,000 more people may have died because of the war. So they've really narrowed it down there.

Not only did this dubious bit of fuzzy math get published, The Lancet pumped it through the peer-review process in six or seven weeks. That happens to be about five or six times faster than a normal academic journal cycle. Why would they do that on the eve of the U.S. presidential election, you ask? No, you probably aren't asking that. You don't need to ask.

UPDATE (10:42am PT): Lancet Methodology Ludicrous

Here is how the "researchers" arrived at their estimated death tolls: they asked locals. That's it. No proof whatsoever. They did not even ask to see death certificates (except in a small fraction of cases): "The research team decided that asking for death certificates in each case, during the interviews, might cause hostility and could put the research team in danger."

If the "researchers" were not suffused with bias they would have immediately seen the inanity of this methodology: opponents of the war and the democratization of Iraq had an incentive to grossly inflate mortality counts, whereas the truly bereaved would not deny their losses.

I used to sit on the peer-review board of an academic journal, and I quit over just such a biased methodology, when a paper was published against my recommendation. The paper relied on student test scores in which the tests had been administered by the students' own family members!!! They had an incentive to inflate those scores, an incentive that would be obvious to any unbiased party. When the journal published it anyway, I walked.

It's sad -- but not surprising -- to see that the Lancet is no better than the rag I walked away from.