Identifying Alcoholics in Population Surveys; A Report on Reliability

Abstract

In 1960-61, 132 alcoholics were identified in a large community health survey conducted in the Washington Heights Health District of New York City, yielding a rate of 19 alcoholics per 1,000 respondents aged over 20 years. To test the reliability of such surveys, 99 of the 132 were located and reinterviewed with their spouses in 1963-64, together with a comparison groups of 343 respondents who had not been identified as alcoholics in the first survey. It was found that 25 of the alcoholics did not repeat their admission of alcoholism in the second survey; and 29 of those first categorized as non-alcoholics now acknowledged a drinking problem. In the second study most of the alcoholics and their nonalcoholic spouses agreed on the presence or absence of alcoholism: 63% of the alcoholics and 79% of the spouses admitted to a drinking problem in the family. Of 90 alcoholics identified in both studies, 9 reported increased drinking since the first study, 21 claimed abstinence, and 28 claimed a decrease in drinking; however, on inspection, only 13 of the latter appeared to be moderate drinkers. Heavy drinking, as measured by the Quantity-Frequency Index, was frequently denied; 54% of those acknowledging a drinking problem denied drinking heavily. It is suggested that this Index is unsuitable for alcoholic populations. A scale constructed to measure attitudes toward drinking failed to identify 47% of the alcoholics.