The primary aims of this study were to empirically test the factor structure of the Children's Eating Attitudes Test (ChEAT) through both exploratory and confirmatory factor analyses and to interpret the factor structure of the ChEAT within the context of a new scoring method. The ChEAT was administered to 728 children in the 2nd through 6th grades (from five schools) at two different time points. Exactly half the students were male and half were female. To the best of our knowledge, this is the first study to empirically test the merits of an alternative 6-point scoring system as compared to the traditionally used 4-point scoring system. With the new scoring procedure, the skewness for all factor scores decreased, which resulted in increased variance in the item scores, as well as the total ChEAT score. Since the internal consistency of two factors in a recently proposed model was not acceptable (<0.60), this model did not adequately fit our data. Thus, exploratory and confirmatory factor analyses were conducted. A 6-factor solution based on a 20-item version was found to best fit the data and have the best internal reliability. The six factors were labeled: 1) overconcern with body size, 2) dieting, 3) food preoccupation, 4) social pressure to gain weight, 5) vomiting, and 6) caloric awareness and control. The obtained factor solution had considerable overlap with the original factor analysis performed on the 26-item Eating Attitudes Test and with the factor structure of the ChEAT reported by previous investigations. Intercorrelations among the factors suggested three higher order constructs. These findings indicate that the ChEAT subscales may be sufficiently stable to allow use in non-clinical samples of children.