Abstract
BACKGROUND: The Eating Disorder Examination-Questionnaire (EDE-Q) is among the most widely used self-report measures of eating disorder (ED) psychopathology. There is a need for brief versions of the EDE-Q that can be used for general assessment and screening purposes. A three-factor 7-item version (EDE-Q7) seems particularly promising but there is a need for more well-powered studies to establish the psychometric properties in both patient and community samples. Moreover, comparing the EDE-Q7 with the full EDE-Q would be beneficial in determining its utility. In the present study, we provide a psychometric comparison between the brief EDE-Q7 and the full EDE-Q in a large sample of both patients and community comparisons. METHODS: We pooled available datasets collected in Norway to amass a large female sample comprising both patients (n = 1954, M(age) = 28 years) and community comparisons (n = 2430, M(age) = 31 years). We investigated the psychometric properties of both versions, including their internal consistency, factor structure, and ability to discriminate between patients and community comparisons. RESULTS: The EDE-Q7 showed similar distributions of scores compared to the full EDE-Q but produced higher scores. Results indicated that the EDE-Q7 have acceptable internal consistency and is adequately able to discriminate between clinical and non-clinical samples. A cut-off threshold of 3.64 was optimal in discriminating between patients and comparisons. We also found support for the three-factor solution for the EDE-Q7, indicating good structural validity. In contrast, we did not find support for the originally proposed four-factor solution of the full EDE-Q. CONCLUSIONS: We find that the brief EDE-Q7 performs close to the full EDE-Q in several respects. Our findings indicate that the brief EDE-Q7 may be a viable alternative to the full EDE-Q in situations where response burden is an issue (e.g., epidemiological studies). However, the EDE-Q7 may hold limited value over the full EDE-Q in clinical settings, due to the small number of items and lack of assessment of behavioral features.