Effects of Head Start On Children's School Readiness: Evidence From a Large National Birth Cohort Study

Background and Purpose: This study examines Head Start's effects on children's school readiness. This issue is important because Head Start is the single largest publicly funded early childhood education program for children's school readiness. This study addresses two common issues found in non-experimental Head Start studies: 1) the selection bias and 2) the lack of a clear reference group. To address selection bias, this study uses several empirical strategies, including OLS regression models with rich controls, state-fixed effects models, and propensity score matching models. To clearly define the reference group, this study compares Head Start participants with non-participants who attended specific types of child care arrangements separately: prekindergarten, other center-based care, other non-parental care, and parental care.

Methods: The data came from the Early Childhood Longitudinal Study - Birth Cohort (ECLS-B), a longitudinal study that tracked from the birth to kindergarten a nationally representative sample of 10,700 children born in the U.S. in 2001. The final sample included approximately 6,950 children (65% of the original sample) belonging to one of five child care groups, measured at the preschool wave: Head Start (17%), pre-kindergarten (15%), other center-based care (39%), other non-parental care (9%), and parental care (19%). Seven outcomes were constructed to measure cognitive development (expressive language, early reading, mathematics), socio-emotional well-being (conduct problems, hyperactivity/inattention, pro-social behaviors), and health status (Body Mass Index) at the kindergarten wave. This study used the propensity score matching method: 1) by conducting probit regressions to predict propensity scores; 2) by matching each Head Start participant with a non-participant who holds the closest propensity score through the one-to-one nearest neighbor matching method with replacement and common support options; and 3) by estimating Head Start's effects through regression-adjusted differences. Also, OLS regressions with state-fixed effects were conducted to serve as a comparison. Multiple Imputation method was adopted to address missing information in some covariates.

Results: Head Start showed no effects on children's cognitive development but negative effects on their socio-emotional well-being and health, compared to non-participants, regardless of specific types of care arrangement. However, substantially differing effects of Head Start were detected when compared to each specific type of care. Head Start presented better cognitive development than other non-parental care or parental care but worse than pre-kindergarten. Also, Head Start consistently showed worse socio-emotional behavior than other center-based care, other non-parental care, or parental care. In addition, Head Start showed worse health status than pre-kindergarten or other center-based care but better than other non-parental care.

Conclusions and Implications: This study shows that clearly defining the reference group offers a better estimate of Head Start effects. Based on the findings, this study suggests that 1) Head Start funding needs to be allocated to target children who must receive other-non parental care or parental care if not attending Head Start, 2) the effort of policy makers to improve the quality of Head Start needs to focus on enhancing Head Start participants' socio-emotional well-being, and 3) the program needs to improve its ability to enhance children's health.