Based on the encouraging findings in part I of our study, and on the relevant research literature, several sociodemographic factors were added to the causal sequence proposed by the distribution of consumption model and corroborated by a statistical technique known as the analysis of linear structural relations (LISREL). The sociodemographic factors were added in two alternative fashions. Although neither approach proved to be significantly superior, the basic causal sequence identified in part I was maintained, and the sociodemographic factors were found to be intimately involved in the process that leads to alcohol-related morbidity and alcohol-related mortality in the general population. The findings are discussed with reference to their implications for prevention via public policy.