poral patterns of consumption within a specified time frame describe important aspects of underage drinking, or drinking by persons of any age for that matter, that are directly linked with the extent and seriousness of potential problems due to that behavior. A number of different measures, therefore, have been developed and are routinely used in epidemiologic surveys to measure patterns of alcohol use.

There are currently no universally accepted measures of alcohol use, and patterns of use, even among researchers. Even a standard dichotomous measure of lifetime use must deal with such methodologically vexing issues as whether and how to exclude alcohol used in religious services. More sophisticated measures of quantity and frequency of alcohol consumption are still confounded by variations in the alcohol content of different types of beverages, and differences in the sizes of standard drinks.

With these limitations in mind, the types of survey instrumentation used to assess alcohol consumption generally may be categorized as follows. “Frequency” measures ask the respondent to report typical drinking frequency over a specified reference period (e.g., a month or year); response categories vary, but may include options such as “never,” “once a month,” “once a week,” and “daily.” “Quantity/frequency” (QF) measures tap both drinking frequency and the average quantity of alcohol consumed on any given occasion, yielding a measure of the total amount of alcohol consumed. Frequency and QF measures may be obtained for alcohol use in general (i.e., any type or alcoholic beverage), or separate responses may be solicited for each specific type of beverage (e.g., beer, wine, hard liquor). “Graduated frequency” measures, a subset of QF, begin with a question concerning the greatest amount of alcohol consumed within a given referent period (and the number of occasions on which this amount of alcohol was consumed), then ask about the number of occasions on which each progressively smaller amount of alcohol was imbibed. Although there is evidence that such measures yield more accurate estimates of drinking patterns and the volume of alcohol consumed (Greenfield, 2000), the number of questions necessary limits the utility of this approach in large, multipurpose surveys. Two other types of alcohol consumption questions, which are less well suited for a survey format, include “short-term recall methods” in which respondents are asked to remember each drink consumed over a several-day period, and “diary methods,” in which participants record their alcohol consumption prospectively. In general, the more detailed the questions, the greater the estimates (Dufour, 1999).

Measures of “heavy” or “binge” drinking, concepts used by the Monitoring the Future (MTF) and National Household Survey on Drug Abuse (NHSDA) studies respectively, also vary in their specific definitions. Heavy or binge drinking is customarily defined as having five or more drinks on any one occasion during a specific period, although for women that number