similar experiences, accumulated over the life course, that help to shape alcohol-related attitudes, beliefs, and drinking behaviors.

Creditable information on the scope, demographic patterns, and trends in underage drinking behaviors provides an important empirical foundation in helping to justify and prioritize needs for policies and programmatic efforts to address this problem and reduce its negative public health consequences. Rigorous analysis of even more detailed epidemiologic data can also contribute to our understanding of the factors that influence this behavior and help suggest more specific strategies for preventing it.¹ Because the purpose of this chapter is to provide a broad overview of underage drinking patterns and trends, rather than examine etiological factors, it does not include complex analyses that simultaneously involve or control for multiple variables. Research based on such strategies, however, will be cited in a number of the more substantively focused chapters in this volume in order to support particular perspectives on underage drinking and approaches for addressing the problem.

Since 1988, and as a direct consequence of the passage of the National Minimum Drinking Age Act in 1984, the minimum legal drinking age for every state in the nation has been 21. In this context, the term “underage drinking” refers to alcohol consumption by persons less than 21 years of age, and is the focus of this volume. Statistics presented in this chapter, therefore, focus primarily on persons under the age of 21. Furthermore, because there appears to be very little alcohol use by persons under the age of 12, and because surveillance data are generally not available for persons younger than 12 anyway, the age range of interest is further delimited by a lower bound of 12 years. Some of the information presented here, however, is based on somewhat different or more restricted age ranges because readily available epidemiological data do not always coincide precisely with this age range, or they pertain to only subsets within this range. Such circumstances are noted accordingly.

In attempting to quantify “drinking” behavior, whether with respect to an individual person or in aggregate, it is clear that no single measure or dimension can adequately capture all the meaningful attributes of this behavior. For example, both the amounts of alcohol consumed and the tem-