March 19, 2002



Iain Murray takes issue with my preceding comments re his refutation of Herold on his site. I have a lot of respect for the guy, and I see where he's coming from. But the simple fact is he's trying to refute a verifiable (if deeply flawed) proposition with a non-verifiable one, an approach that can never be conclusive. Murray's also being a touch disingenous when he writes:

We [Stats] do not produce new data ourselves, except on rare occasions when we have a specific grant for the project.

Murray says he has done no independent evaluation of the number of Afghan civilian casualties. But he also stands by his clincher in his opinion piece that his own thorough review concluded that 650 of the deaths Herold found are "in any way reliably reported." (Those are not mutually contradictory statements: a review of existing data and the generation of new figures are different things, of course.) But that still doesn't explain why, if Murray had any faith in the methodology behind his review of the Herold figures, that he's not willing to expose it to its own scrutiny. And if he doesn't want to share it, for whatever reason, why he thought referring to it would make his argument any the stronger. If anything, that unsupported reference considerably weakens his piece, and ventures very close to idle self-promotion for Murray and Stats.

Look at it this way. If Marc Herold had published his 3,000-plus figure for Afghan casualties, and not offered all his data for Murray and others to challenge, Murray (and I) would have rightly torn him to pieces for not having the guts to expose his data to scrutiny. But now Murray makes a claim, backed by a study he's already completed, but he's content to keep it securely in his pocket, instead asking the rest of the world to "just trust us."

Posted by BruceR at 02:07 PM