Evaluating an abbreviated three-factor version of the Eating Disorder Examination Questionnaire in three samples

A brief, three-factor structure for the Eating Disorder Examination-Questionnaire (EDE-Q) has been identified that has stronger psychometric properties than the original four-factor EDE-Q structure. However, there is a need for independent replication of the brief version in different samples, and examination of how comparable the brief version is in women and men of diverse ages. This study compared factor structure fit and other psychometric properties between the original and brief versions in three independent samples of adults to evaluate the reliability and validity of this brief version. It also examined measurement invariance across gender in a middle-aged community sample. Participants included university women (N = 659), university women and men (N = 358), and crowdsourced community women and men (N = 544). Across samples, compared to the original version the brief version demonstrated greater internal consistency, somewhat less item overlap, and superior model fit in confirmatory factor analyses, although some fit indices fell below recommended cut-offs. Separate confirmatory factor analyses of the brief version in community women and men suggested the brief version may better fit community men than women, and measurement invariance across gender for the brief version was not supported for the community sample. This independent replication supports the factorial validity of the brief version in three samples of students and crowdsourced community adults, and highlights the importance of further examination of the psychometric properties of this brief three-factor version across gender and age categories, among others sample characteristics. •The brief EDE-Q demonstrated superior model fit compared to the original version.•The brief EDE-Q had greater internal consistency and somewhat less item overlap.•Measurement invariance was not supported across gender in a community sample.•Importance of further examination of the brief EDE-Q across samples was highlighted.