William C. Kerr, a scientist with the Alcohol Research Group, lead a study that found that American drinkers who measure their usual alcohol consumption in the home environment have considerable variation in drink size, particularly for spirits and wine.
"Individual drinkers should be concerned with varying drink alcohol content because the consumption of non-standard drinks affects their ability to keep track of how much alcohol they have consumed and therefore their ability to conform to safe drinking guidelines and driving laws," said Kerr, in a news release about the study.
"Without valid and reliable measures of how much alcohol is being consumed by a population, the ability to assess risk is compromised," said Lorraine Midanik, an affiliate senior scientist with the Alcohol Research Group. "Population estimates of alcohol use derived from self-reports are used to determine risk for various adverse health and social outcomes. Population estimates of alcohol use also help us to understand trends in the U.S. population. That is, 'are people drinking more, less or about the same alcohol over time?,' 'how are these trends related to morbidity and mortality?,' and 'what does this mean for policies that may affect alcohol consumption?'"
Kerr said many countries have standard drink sizes that differ from the 0.6 ounce standard. "Most of these, such as the United Kingdom standard of eight to 10 grams of alcohol are smaller than the U.S. standard," he said. "However, in no case are these standards based on studies of what drinkers actually consume. In fact, this is the first study to utilize beaker-measured drink sizes and brand-level percentage alcohol by volume information in a national sample in the U.S."
Simplifying the Message"The mean drink alcohol content (in the survey) was 0.67 ounces, which is 11.7 percent larger than the 0.6 ounce standard drink," said Kerr. "A single-drink standard is helpful for simplifying messages about safe drinking limits and driving-under-the-influence limits as compared to having different standards for each beverage."
"Wine and spirits drinks are on average larger and more variable than beer drinks," said Kerr, "with spirits drinks averaging nearly 50 percent more alcohol than the U.S. standard drink."
Implications for Research"Clearly there is much variation in what constitutes 'a drink' in the U.S. population," said Midanik. "The size of a drink may vary dramatically, as well as the alcohol content of the alcohol consumed. These findings should stress to researchers that obtaining information on beverage use must include specific alcohol beverage type and some measure of the size of drinks."
"This discrepancy is larger for women," Kerr said, "because a larger proportion of their drinks are wine and spirits. Also, the difference in monthly alcohol volume is larger than the difference in mean drink size because heavier drinkers appear to choose high-alcohol-content drinks on average."
Underestimating Alcohol Use"The real-world implications of these findings are that we are continuously underestimating alcohol use in the U.S. and within specific populations with our traditional survey methodology," said Midanik. "This is highly problematic given that policy, treatment and prevention efforts are based on national surveys which provide basic epidemiologic data on alcohol use and alcohol-related problems in the population. These findings also have major implications for researchers, and will hopefully provide additional refinements to our survey questions in order to provide better estimates of alcohol consumption."
"In summary," said Kerr, "our findings suggest that many readers and their friends and family are pouring themselves larger-than-standard drinks. Given that safe-drink messages are based on a specific amount of alcohol in a drink, 0.6 ounces, they might want to become more aware of the percentage of alcohol in the brands of beer, wine and spirits they choose, perhaps adjusting their drink size accordingly. At minimum, they should be aware of larger-than-standard drinks."
Source: Results of the study were published in the November 2005 issue of Alcoholism: Clinical & Experimental Research.