Abstract
Estimates of alcoholism prevalence are needed for propaganda purposes, for planning and for ecological and epidemiological studies designed to understand etiology. For all those purposes, nearly all current estimations are founded on the Jellinek Estimation formula., -which is based on annual liver cirrhosis death rates. Neglecting empirical difficulties (e.g. sampling) it is shown by an elementary symbolical-logical analysis that credible estimates cannot be reached by the methods recommended, that the systematic errors introduced are sizeable, and that it is im probable that the true rates are even linear transformations of present estimated rates. It is suggested both that attempts be made to correct the blatant errors (for short-run purposes) and to develop a new method of estimation (for long-run use).

This publication has 5 references indexed in Scilit: