Abstract
The yearly sampling stability of alcoholism prevalence statistics [for humans] was investigated. Consumption data were preferable to mortality data for areas with populations of < 1 million.