Two procedures to reduce response bias in reports of alcohol consumption

Standard instructions concerning the extent of confidentiality and a retrospective diary procedure were examined for their effects on reports of alcohol consumption. A total of 156 individuals attending a health fair at a southern university participated in an alcohol survey. Subjects were randomly distributed an alcohol survey that presented a 7-day and a 28-day quantity-frequency (Q-F) measure and a retrospective diary measure. Subjects were also randomly distributed an informed consent sheet that stated that their responses would remain completely anonymous and confidential or one that stated that their responses would be kept completely confidential and asked them to identify significant others to verify their responses. A significantly greater proportion (p = .0003) of subjects receiving instructions of confidentiality with verification did not sign the consent form or did not identify at least one significant other. No differences were found across types of instructions on measures of alcohol consumption and willingness to report alcohol, cigarette and marijuana use. The retrospective diary resulted in significantly greater reports of the number of drinking days per week (p = .0004) and number of drinks per week (p = .02) than did 7-day and 28-day Q-F measures. The diary measure also classified a greater percentage of drinkers as heavy drinkers, whereas the 28-day Q-F measure classified a greater percentage of subjects as drinkers.