A detailed comparison was made of two methods for assessing the features of eating disorders. An investigator-based interview was compared with a self-report questionnaire based directly on that interview. A number of important discrepancies emerged. Although the two measures performed similarly with respect to the assessment of unambiguous behavioral features such as self-induced vomiting and dieting, the self-report questionnaire generated higher scores than the interview when assessing more complex features such as binge eating and concerns about shape. Both methods underestimated body weight.