Examining the factor structure and measurement invariance of the online-administered Eating Disorder Examination-Questionnaire and the Eating Attitudes Test-26 in young and middle-aged women.
Article
Bazo Perez, Maria, de Carvalho, Pedro Henrique Berbert, Frazier, Leslie D. (2025). Examining the factor structure and measurement invariance of the online-administered Eating Disorder Examination-Questionnaire and the Eating Attitudes Test-26 in young and middle-aged women.
. EATING AND WEIGHT DISORDERS-STUDIES ON ANOREXIA BULIMIA AND OBESITY, 10.1007/s40519-025-01802-8
Bazo Perez, Maria, de Carvalho, Pedro Henrique Berbert, Frazier, Leslie D. (2025). Examining the factor structure and measurement invariance of the online-administered Eating Disorder Examination-Questionnaire and the Eating Attitudes Test-26 in young and middle-aged women.
. EATING AND WEIGHT DISORDERS-STUDIES ON ANOREXIA BULIMIA AND OBESITY, 10.1007/s40519-025-01802-8
Widely used eating disorder (ED) measures, such as the Eating Disorder Examination-Questionnaire (EDE-Q) or the Eating Attitudes Test-26 (EAT-26), were originally developed and standardized in young White women, leading to poor performance, unclear factor structures, and inconsistent measurement invariance across diverse groups. As ED prevalence rises among middle-aged women, the need for age-appropriate and psychometrically sound assessment tools has become increasingly important. This study evaluated the factor structure, measurement invariance, and internal consistency of the EDE-Q and EAT-26 when administered online across two developmentally relevant age groups: emerging adults and middle-aged women.
Method
A sample of 829 women from across the U.S. (emerging adults: 419; middle-aged: 410) completed the EDE-Q and EAT-26 through an online survey platform. We tested the original factor structures and two alternative models for each measure through confirmatory factor analysis. Measurement invariance analyses were conducted on good-fitting models.
Results
The original EDE-Q model failed to converge, while the original EAT-26 model demonstrated poor fit. The alternative factor models-Grilo et al.'s (2013) EDE-Q model, and Bazo Perez et al.'s (2023) EAT-26 model-demonstrated best fit and measurement invariance across both age groups. The EDE-Q subscales exhibited good internal consistency, while the EAT-26 showed acceptable to good internal consistency.
Conclusion
These findings emphasize the need for developmentally sensitive tools to improve diagnostic accuracy, early detection, and treatment of EDs across the lifespan. Because the factor structure and measurement invariance results reflect online administration, they should be interpreted within this context and motivate continued evaluation of these instruments across administration formats. Addressing a critical gap in ED research and clinical practice, this work underscores the need to refine ED assessment methods, to ensure equitable, accurate, and developmentally appropriate identification of ED risk in women beyond early adulthood.
