Background and purpose: The Minimal Eating Observation Form-Version II (MEOF-II) is a screening and research tool for eating difficulties. Here we assess the psychometric properties of the tool. Methods: Cross-sectional MEOF-II data, n = 5,956 persons (nursing home residents [n = 3,087]; in-hospital patients [n = 1,490]; mean age 76.9; 61% women). Data was analyzed taking account of the dichotomous nature of item-level data, that is, based on tetrachoric correlations. Results: The MEOF-II items (n = 9) belonged to one higher order unidimensional factor (61.7% explained variance, omega 0.92), and three lower level factors, with three items each (92.9% explained variance, omega 0.83-0.88): deglutition; ingestion; and energy/appetite. Conclusion: The study provides support for the reliability and validity of the MEOF-II total scale and subscales, and the findings are in line with previous studies.