Quality care and education are fundamental components of the services provided by Early Head Start (EHS) programs serving children from birth to age 3. Research evidence indicates that child care quality is associated with child outcomes. However, most studies on quality of early care and education have been done with preschool classrooms. In this paper, we describe the overall quality of Early Head Start classrooms and examine how staff and program characteristics relate to classroom quality in EHS and whether classroom quality associates with child outcomes, using data from the Early Head Start Family and Child Experiences Survey (Baby FACES). Baby FACES is a longitudinal study of a nationally representative sample of EHS programs following two cohorts of children (newborns and 1-year-olds) who were enrolled in spring 2009 until age 3 or they left the program.

To measure classroom quality in center-based programs, we used the Classroom Assessment Scoring System-Toddler (CLASS-T), a new measure of effective teacher-child interactions with children ages 15 to 36 months, to observe classrooms when children were 2 and 3 years old. Because some of the teachers were teaching both cohorts of children and were therefore observed for 3 years, we combine the data for the two cohorts to examine the pattern of classroom quality over time and the predictors of overall classroom quality.

We find that Early Head Start children are in classrooms with ratings in the mid-range of quality (scores of 3 to 5 out of 7) in the domains of Emotional and Behavioral Support and Engaged Support for Learning. Scores in Engaged Support for Learning are lower, on average, than those observed in Emotional and Behavioral Support. These findings mirror those in studies with older preschool children and suggest that providers face challenges in their attempts to offer high-quality instructional support to children.

We construct 3-level HLM models to examine whether classroom quality changed over 3 years and what teacher and program characteristics predict overall classroom quality. Results show that the mean scores for Engaged Support for Learning decreased from 2010 to 2012, but the means for Emotional and Behavioral Support remained consistent over time. Among staff characteristics examined in the models, teacher education and teacher-parent relationships were positively associated with classroom quality, while teacher depressive symptoms were negatively associated with classroom quality. None of the program characteristics in the models were related to classroom quality.

The analyses linking classroom quality and child outcomes demonstrate some evidence of threshold effects. We find that the associations between quality and some child outcomes were stronger in classrooms scoring in the higher quality range than in classrooms in the lower quality range. In general, these findings are consistent with the research literature that examines threshold effects in the relationships between classroom quality and preschool children’s outcomes.