Tragedy doesn’t scale

No one knows how many people—civilians or combatants—were killed in the battle for Tskhinvali, and both self-serving and independent estimates have varied widely since the conflict escalated in early August. The BBC now reports that

The Russian prosecutor’s office is investigating more than 300 possible cases of civilians killed by the Georgian military.

Some of those may be Ossetian paramilitaries, but Human Rights Watch believes the figure of 300-400 civilians is a “useful starting point”.

That would represent more than 1% of the population of Tskhinvali—the equivalent of 70,000 deaths in London.

This sort of “equivalence” is often cited, to convey the impact of a loss on the local population, whether in Iraq,

Thus, violence killed 300 Iraqis last week, the equivalent proportionately of 3,300 Americans,

or in Israel,

Israel’s generals warn that if the Arabs strike first, tiny Israel could suffer 40,000 dead (the American equivalent of nearly three million).

This may be a useful rhetorical device in some narrow demographic sense, but how difficult is it to grasp “one percent”? And there’s a difference between distributed, ongoing casualties—as in Iraq at the time of that article—and what is essentially a single incident. In the former case, the comparison to the U.S. gives a sense of the incidence of violent death, if it were to strike widely and frequently. People often understand frequency-based explanations better than rate-based.

This analogy is likely to sensationalize an already grave matter. Individual lives are not worth more because the victims come from a small community. A car crash in Liechtenstein that kills four people leaves four people dead—it is not equivalent to an earthquake in China that kills 150,000, though each is a comparable share of the total population. 300 dead (or whatever the true figure turns out to be) is terrible enough, but the moral culpability is not the same as for whomever might kill 70,000 civilians in London. On the other hand, thinking about 70,000 dead in London might harden the BBC listener’s heart; one dead man is death, two million are only a statistic (Remarquenot Stalin.)

Update: It seems that Jordan Ellenberg took up this issue a few years ago. I was a math teammate of his; he was much younger and much smarter than I, so I’m not about to take issue with his analysis.