The Early Childhood Environment Rating Scale - Third Edition (ECERS-3) is the latest version of one of the most widely used observational tools for assessing the quality of classrooms serving preschool-aged children. This study was the first assessment of its factor structure and validity, an important step given its widespread use. An ECERS-3 observation was conducted in 1063 preschool classrooms in three states. In a subset of those classrooms (n = 119), Classroom Assessment Scoring System - Pre-K (CLASS Pre-K) and child assessment data were also collected. Analyses of the ECERS-3 suggested that a single factor does not adequately capture item variability. Of the solutions tested, the four-factor (Learning Opportunities, Gross Motor, Teacher Interactions, and Math Activities) provided the best combination of statistical support and theoretical utility. In general, the ECERS-3 Total Score and the four factors were moderately correlated with the three domains of the CLASS Pre-K. ECERS-3 Total Score, Learning Opportunities, and Teacher Interactions were positively related to growth in executive function, as were all three domains of the CLASS Pre-K. However, all significant associations were small, and most tested associations between ECERS-3 scores and children's growth, and between CLASS Pre-K and children's growth, were not significant. Results are discussed in terms of their implications for measuring preschool quality. (author abstract)