Recorded presentation

Yu Ye, MA, Thomas K. Greenfield, PhD, Jason Bond, PhD, and William C. Kerr, PhD. Alcohol Research Group, 2000 Hearst Avenue, Berkeley, CA 94709, 510 642-5208, yye@arg.org

Heavy drinking episode, measured by the frequency of consuming five drinks or more in a day, has become a standard question in public health surveys, in which an open-ended format is commonly used. Using the 2000 US National Alcohol Survey, we compare the 12-month open-ended (OE) 5+ with a similar but categorized measure derived from the closed-ended Graduate Frequency (GF) questions among current drinkers (n=4630). The initial agreement between the two measures was assessed by creating ordered categories for each variable. To evaluate the accuracy of the heavy drinking measures, the two variables, as well as some derivatives (e.g. the average, maximum, etc), were used to predict alcohol-related tangible consequence and dependence criteria, controlling the demographic variables. A J-test for non-nested models was conducted on the logistic regressions to compare the prediction performance. Moderate agreement was observed between the two categorical measures. While OE 5+ obtained a higher rate of any 5+ for less heavier drinkers, GF 5+ reported a higher number of heavier drinkers. GF 5+ also outperformed OE 5+ in predicting consequence and dependence criteria. Of all measures evaluated, the maximum of OE and GF 5+ performed the best. We conclude that open-ended questions might not be as accurate as the closed-ended Graduate Frequency questions, at least for heavier drinkers and a 12-month reference period. Consistent with psychometric theory, a better measure of frequency of heaving drinking may be obtained by combining multiple measures of heavy drinking.

Learning Objectives: At the end of this presentation, the participant will be able to

• Construct the heavy drinking question formats in surveys.
• Discuss the agreement between different measures and their accuracy.
• Develop the more accurate measures of heavy drinking.

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.

Recorded presentation