[mjburgess]

> This seems... peculiar?

It's also just false. It applies to normally distributed quantities whose average is centred on the same scale the population is measured on.

However it, necessarily, doesn't apply, to power-law distributions (,... wealth, popularity, ...) -- and many others.