The Eating Attitudes Test-40 (EAT-40; Garner & Garfinkel, 1979) is one of the most widely used measures in the field of eating disorders (ED). The factor structure of the EAT-40, as well as the optimal cut-off score to identify subjects with ED, are subjects of debate. Both controversial issues are addressed in the present study. Participants were 95 clinical females meeting DSM-IV-R criteria for ED and 89 females without ED. The results supported a unidimensional structure of the EAT-40 items scores. The general factor accounted for a high percentage (50.63%) of the variance in EAT-40 total scores. The questionnaire proved to have good internal consistency and test-retest reliability. Clinical participants displayed higher mean scores than normal subjects in the EAT-40. Further, participants meeting DSM-IV-R criteria for ED differed significantly from "symptomatic" and "asymptomatic" participants. Correlations with the BULIT-R and the EDI-II scores supported the convergent validity of the questionnaire. The EAT-40 also demonstrated good specificity (94.38%) and sensitivity (93.68%) to detect ED when a cut-off score of 27 was used to discriminate between subjects with and without ED. The implications of these findings for the conceptualization and the assessment of ED are discussed.