Regularity in the Distribution of Alcohol Consumption

Abstract

The first association of the word deviance is usually not to high blood pressure, overweight and excessive salt consumption, as in the title of Rose and Day (1990). In sociology, deviance is a social construct, referring to behaviours that society regards as undesirable, abnormal, or antisocial (directed aggressively towards or indifferently away from other people). In most cases, deviant behaviour can be distinguished in a qualitative way from 'normal' behaviour, in the sense that people easily recognize or identify it. Criminal conduct, illicit drug use and prostitution, but also divorce and suicide, are examples of deviant behaviour as can be found in sociological publications. Generally, it is negatively valued and regarded destructive for the individual or for society. Often it is rooted in the subculture. Alcoholism or excessive drinking has much in common with this meaning of deviance. However, since most adults in western society drink alcoholic beverages in one way or another, recognition of deviant drinking behaviour is not as easy as in drug use or criminal behaviour (Robins, 1975). The question of what is a normal or social drinking habit and what constitutes abuse or alcoholism has bothered western society for centuries. In modern times, two main views on this issue have dominated the field. One asserts a qualitative difference between normal and pathological drinking behaviour, with little or no association between a general level of consumption and the number of alcoholics or the number of problems in society caused by excessive drinking (the socalled disease model). The other view, an epidemiological model, stresses that the whole population is at risk of experiencing the negative consequences of drinking, owing to an alleged regularity in distribution of consumption and a resulting close connection between mean consumption and excessive drinking. This latter singlepopulation model will be the focus of this paper. First, a brief description of the single-distribution model is given, together with problems of measurement and estimation. At the end some of the implications for prevention are discussed. As an explanation for the regional variations in liver cirrhosis mortality in France, the epidemiologist Ledermann (1956, 1964) proposed a close connection between average consumption and excessive use, and he set out to model this relationship statistically. Using several data sets from France and abroad, he arrived at a lognormal distribution model with a fixed relation between the mean and dispersion parameter. Knowledge of the total volume allowed for the estimation of prevalence of