A validation study was performed in order to compare the observed versus the self-reported food consumption in anorectics (n = 9), anorectic bulimics (n = 6) and bulimics (n = 10). The observed meals (breakfast, lunch and/or dinner) were served in a hospital setting. Foods on plates were unobtrusively weighed before and after serving to each subject, and the foods consumed were compared with data from a 24-hour recall the subsequent day. Data analyses via a nutrient database comprised amounts of food items from seven food groups as well as energy, macro nutrients, vitamin C, calcium and iron. Analyses of recalled intakes showed consistently high correlations with observed intakes, suggesting that the 24-hour recall gives valid results, and could thus be suitable for this category of patients. This notion contrasts with previous statements, although no validation studies have been undertaken in order to support this view