Participation in everyday routines (e.g., mealtimes, bedtime, dressing) is a key context for early development and a central focus of early childhood intervention. This study examined the internal structure, reliability and construct validity of the Measure of Engagement, Independence and Social Relationships for children aged 3–5 years (MEISR 3–5) within home routines. The sample included 705 children aged 36 to 72 months (M = 51.01, SD = 10.07) receiving early childhood intervention services across Spain. A confirmatory factor analysis based on McWilliam's theoretical functional domains—engagement, independence and social relationships—showed acceptable global model fit and very high internal consistency. However, convergent validity was only partial, and discriminant validity was not supported, indicating substantial overlap amongst domains. To further examine the dimensionality of the instrument, parcel-based analyses were conducted in a subsample (n = 553) and compared correlated three-factor, two-factor, bifactor, second-order and unidimensional models. All models showed equivalent and excellent fit. Bifactor indices (ωh ≈ 0.98; ECV > 0.90) indicated that most reliable and common variance was attributable to a dominant general participation factor, with minimal unique variance remaining for the specific domains. These findings suggest that the MEISR 3–5 functions as an essentially unidimensional measure at the psychometric level. The total score appears to provide the most robust quantitative indicator of children's participation in home routines, whereas domain-level scores are better interpreted as clinically descriptive dimensions that can inform goal setting, intervention planning, and progress monitoring in routine-based practise.