Abstract
At a reinterview after 18 years, 67 men overestimated their past alcohol consumption by an average of 76%. The original consumption was the best regressor of the recalled consumption, but the magnitude and direction of error in remembering the original consumption varied considerably.