Objective: This study aimed to provide an empirical foundation for selecting optimal models of the Eating Disorder Examination Questionnaire (EDE-Q) full-scale and short forms in Japanese female clinical and nonclinical populations.
Method: The sample comprised 526 female patients with eating disorders and 744 female controls. The analysis systematically compared psychometric properties, including factor structures, internal consistency, and discriminative capabilities across five full-scale models and three short-form models of the EDE-Q. Furthermore, factorial estimates were evaluated across diagnostic subgroups: anorexia nervosa (AN), bulimia nervosa (BN), and binge-eating disorder (BED).
Results: Confirmatory factor analysis (CFA) did not support the proposed factor structure for any of the EDE-Q models, except for the 7-item (EDE-Q7) and 9-item (EDE-Q9) forms. The three-factor EDE-Q7 model, comprising dietary restraint, shape/weight overvaluation, and body dissatisfaction, was validated in the clinical sample but not in the nonclinical sample. Conversely, the four-factor EDE-Q9 model, comprising dietary restraint, shape/weight overvaluation, body dissatisfaction, and preoccupation, demonstrated structural validity across both clinical and nonclinical samples. CFA results indicated optimal model fit for the EDE-Q9 in the AN, BN, and BED groups, whereas for the EDE-Q7 in the BN and BED groups. Both short forms exhibited robust internal consistency and effectively differentiated between clinical and nonclinical samples.
Conclusions: The findings indicate that specific short forms of the EDE-Q offer viable alternatives in time-constrained settings, exhibiting more consistent structural validity compared to the full-scale version.
Keywords: Confirmatory factor analysis; Discriminative ability; Eating disorder; Eating disorder examination questionnaire; ROC curve analysis.
Copyright © 2025 Elsevier Ltd. All rights reserved.