Restrained eaters have been reported to overeat following a high caloric preload, a phenomenon referred to as the disinhibition effect. However this effect has not been found when subjects were classified by the restraint subscales of the Three-Factor Eating Questionnaire (TFEQ; Stunkard & Messick, 1985) or the Dutch Eating Behaviour Questionnaire (van Strien et al., 1986). The present study investigates the disinhibition effect in 133 normal-weight young women, using a two-factorial classification including the TFEQ-restraint and the TFEQ-disinhibition scale. The subjects were requested to consume ice-cream ad libitum during a taste test following a 200-ml milkshake preload or without preload. The results show that the behavioural disinhibition effect occurs only in subjects with simultaneous high scores on both subscales. In addition, subjects with high disinhibition scores consumed more ice-cream than low disinhibition subjects irrespective of their degree of restraint. While subjects with a more rigid control of eating behaviour did not show a difference in the amount of ice-cream consumed with or without preload, subjects with a more flexible control of eating behaviour reduced their intake following the preload condition. With regard to the Revised Restraint Scale (RRS Herman & Polivy, 1980) multiple regression results show that high RRS scores may be due to either higher TFEQ-restraint or higher TFEQ-disinhibition scores. The interpretation of the results favours the renaming of the TFEQ-disinhibition scale to "susceptibility to eating problems" because high scores on this scale indicate overeating in a variety of situations without requiring prior inhibition i.e. dietary restraint. It is supposed that high susceptibility to eating problems may be caused by rigid control of eating behaviour, whereas flexible control of eating behaviour may be a less problematic strategy of long-term weight control.