Abstract
The problem of explaining the distribution of alcohol consumption in a population is discussed with particular reference to the log-normal frequency distribution of alcohol consumption and the claims associated with it. Criticisms directed at this application of the log-normal distribution are presented, as is Ole-Jørgen Skog's defense of the log-normal hypothesis. A nation-wide survey of alcohol consumption by 1278 New Zealand 14- to 17-yr.-olds is described. In this sample, representing homogeneous substrata of the New Zealand population, the distribution of alcohol consumption is not close to log-normality in the high consumption segment of the distribution for males, females, and the total sample. However, some analysts would probably interpret our results as borderline or even approximate log-normal distributions. The significance of these results for the log-normal hypothesis and Skog's theoretical position is outlined. Important features of alcohol consumption among 14- to 17-yr.-olds in New Zealand are discussed.