Nature of questionnaire options affects estimates of alcohol intake

To estimate the degree of underreporting of alcohol consumption caused by questionnaire options for the frequency and quantity of drinking, two questionnaires (one oriented toward light alcohol consumption and the other toward heavy consumption) were administered to each of 86 patients (13 women) admitted for withdrawal treatment. The estimated mean daily alcohol intake was 137 g of absolute alcohol as measured by the "light" questionnaire and 302 g as measured by the "heavy" one. The Spearman correlation coefficient between individual mean intakes recorded by the two questionnaires was .58, indicating some influence on the rank ordering of drinkers. The intake estimate based on the heavy questionnaire correlated better with the number of drinks needed to overcome the subjective feelings of hangover than did the estimate based on the light questionnaire. A good deal of the underreporting commonly found in general population surveys might be related to the nature of questionnaire options. More detailed options for heavy intake might reveal previously unnoticed correlations between alcohol intake and various health hazards.