An important consideration in determining the validity of an observational assessment measure for young children is the variability attributed to the child versus that ascribed to the assessor or to some other factor such as classroom context. The "Teaching Strategies GOLD"® assessment system was used to elicit teacher ratings of a national sample of 21,592 children (age 12-51 months). Teacher ratings of child development and learning were associated in expected directions with both child demographic characteristics and classroom composition variables. Children with disabilities started behind their typically developing peers and grew slower, girls showed an advantage in some areas over boys, and English language learners (ELLs) were rated lower at the beginning of the year and showed some faster rates of growth than their native English-speaking peers.