Establishing the measurement invariance of the Eating Disorder Inventory across Hispanic White, non-Hispanic Asian, non-Hispanic Black or African American, and non-Hispanic White adults.

= 2,931) were Hispanic White (7%), non-Hispanic Asian (18%), non-Hispanic Black or African American (7%), and non-Hispanic White (68%) college students (67% female) recruited in a cohort-based epidemiological, longitudinal study of health and eating patterns. First, multigroup confirmatory factor analysis examined whether the EDI performed comparably across ethnoracial groups in the full sample. Next, the same approach was followed in sensitivity analyses within each sex. Evaluation of changes in comparative fit index indicated that full metric invariance, scalar invariance, and uniqueness were supported across ethnoracial groups in the full sample and in women (change in comparative fit index ≤ .01). Only partial measurement invariance was supported in analyses of non-Hispanic Asian and non-Hispanic White men. Analyses support use of the EDI in ethnoracially diverse samples including women and men or only women. Limited score variance in men suggests the need for tests of measurement invariance in larger samples. (PsycInfo Database Record (c) 2026 APA, all rights reserved).