1 Department of Psychology, The Ohio State University, Columbus, OH, USA. Electronic address: tylka.2@osu.edu.
2 Laboratory of Cyberpsychology, Department of Psychoeducation and Psychology, Université Du Québec en Outaouais, Saint-Jérôme, Canada; Substantive-Methodological Synergy Research Laboratory, Department of Psychology, Concordia University, Montreal, Canada.
3 School of Psychology, Deakin University, Geelong, Victoria, Australia.
4 Department of Psychology, Michigan State University, East Lansing, MI, USA.
5 School of Psychology, Sport, and Sensory Sciences, Anglia Ruskin University, Cambridge, United Kingdom; Centre for Psychological Medicine, Perdana University, Kuala Lumpur, Malaysia.

Intuitive eating has been found to protect against disordered eating and preserve well-being. Yet, there are methodological (length), conceptual (inconsideration of medical, value-based, and access-related reasons for food consumption), and psychometric (item wording) concerns with its most common measure, the Intuitive Eating Scale-2 (IES-2). To address these concerns, we developed the IES-3 and investigated its psychometric properties with U.S. community adults. Across three online studies, we evaluated the IES-3's factorial validity using exploratory factor analysis (EFA; Study 1; N = 957; M_age = 36.30), as well as confirmatory factor analysis (CFA), exploratory structural equation modeling (ESEM), bifactor-CFA, and bifactor-ESEM (Study 2; N = 1152; M_age = 40.95), and cross-validated the optimal model (Study 3; N = 884; M_age = 38.54). We examined measurement invariance across samples and time, differential item functioning (age, body mass index [BMI], gender), composite reliability, and validity. Study 1 revealed a 12-item, 4-factor structure (unconditional permission to eat, eating for physical reasons, reliance on hunger and satiety cues, body-food choice congruence). In Study 2, a bifactor-ESEM model with a global intuitive eating factor and four specific factors best fit the data, which was temporally stable across three weeks. This model also had good fit in Study 3 and, across Studies 2 and 3, and was fully invariant and lacked measurement bias in terms of age, gender, and BMI. Associations between latent IES-3 factors and age, gender, and BMI were invariant across Studies 2 and 3. Composite reliability and validity (relationships with disordered eating, embodiment, body image, well-being, and distress; negligible relationship with impression management) of the retained model were also supported. The 12-item IES-3 demonstrates strong psychometric properties in U.S. community adults. Research is now needed using the IES-3 in other cultural contexts and social identity groups.