Choosing High Quality Preschool: Estimating the Causal Effects for Children's Development

The association between high quality preschool experiences—particularly the quality of teacher-child interactions—and positive cognitive and social-emotional outcomes for children has been firmly established through a large body of observational research (Burchinal et al., 2008; Frede, 1995; Howes, Phillips, & Whitebook, 1992; NICHD, 2002; Pianta, Barnett, Burchinal, & Thornburg, 2009; Zaslow et al.). However, the field has not yet been able to overcome issues of selection bias in families’ preschool choices to isolate the causal impacts of high quality care (Pianta et al., 2009).

 The current study uses multi-level propensity score matching methods to explore the causal impact of a family’s choice to enroll their child in a preschool setting characterized by high quality teacher-child interactions on the child’s cognitive and social-emotional development. A sample of 2430 3 and 4 year old children enrolled in formal center and home based care during the 2002-2003 school year were drawn from the Head Start Impact Study (HSIS; HHS, 2010), a nationally representative randomized controlled trial of the impacts of Head Start.  

 In the fall and spring, children in the HSIS were given an extensive battery of cognitive assessments (including the PPVT and the Woodcock Johnson III) and parents and teachers rated children’s social-emotional skills on several measures including a modified version of the Student-Teacher Relationship Scale (Pianta, 1992). Classroom quality was assessed using the Early Childhood Environment Rating Scale (Harms, Clifford, & Cryer, 1998) and the Arnett Caregiver Interaction Scale (Arnett, 1989). Extensive measurement work was done to leverage the additional precision and salience afforded by multiple measures. Resulting domain scores serve as our measures of both quality and child outcomes.

 A multi-level logistic regression and matching with replacement was used to identify a counterfactual group of children enrolled in low quality care who, based on a set of baseline characteristics, were similarly likely to enroll in high quality care as those who actually did. The propensity scores resulting from these models were then used as weights in regression analysis with covariates to predict the impact of enrolling in high quality care on children’s cognitive and social-emotional development over the course of their preschool year.

 Results address the question of whether children whose families’ choose to enroll them in settings characterized by high quality teacher-child interactions make greater cognitive and social-emotional gains than they would have had they been enrolled in lower quality care. Implications for interventions designed to support families in making choices about child care will be discussed.