It has been suggested that future research into the adverse effects of binge drinking in pregnancy should take into account the time of binge drinking, and a method has been proposed. In a representative sample of 1,311 pregnant Danish women in 2000, the agreement between two different measures of binge drinking during the first half of pregnancy obtained from interviews and questionnaires was assessed. Compared with self-administered questionnaires, the interview provided an overall higher response rate and a higher response rate to the specific questions about binge drinking. Furthermore, a higher proportion of women admitted to binge episodes, and the internal consistency was better in the interviews compared with the questionnaires. The percent agreement between the methods ranged between 76% and 100% irrespective of the definition of binge drinking. Self-administered questionnaires are generally easier and cheaper to administer than interviews and, for descriptive purposes, the questionnaires may be sufficient to get an overall impression of the distribution of binge drinking. However, if the goal is to obtain data on the association of binge episodes during specific weeks of pregnancy and adverse pregnancy outcomes, the present results seem to suggest that personal interviews should be conducted if at all possible, particularly because of high internal consistency.