Short-term and long-term effects of Head Start: A revisit using the Early Childhood Longitudinal Study

Includes bibliographical references (pages 101-106)This thesis aims to assess both the short-term and long-term effects of Head Start on children’s test scores, special education placement and grade retention using the Early Childhood Longitudinal Study Kindergarten Cohort of 1998-99 (ECLS-K), a recent national survey data. In addition to OLS regression, quasi-experimental methods such as fixed school effects regression, propensity score matching, random effects panel analysis and a recursive bivariate probit are performed to eliminate the bias caused by non-random selection into Head Start programs and the heterogeneity among Head Start centers and schools. The ECLS-K data show Head Start participants, when compared with other children, are children with disadvantaged family backgrounds. They typically have low test scores and high probabilities of receiving special education and repeating grades. However, the data also show race heterogeneities. For example, the gap between white Head Start participants and other white children expands over time while this is not observed for black and Hispanic children. My empirical findings show that in the short term, black students benefit from Head Start on the general knowledge score and Hispanic students benefit from Head Start on all scores including reading, math and general knowledge. In the long term, I find that Head Start reduces the probability for black students of repeating grades and being placed in special education. It also reduces the probability for Hispanic children of special education placement. However, I find that Head Start does not have a long-lasting positive effect on test scores nor do white students receive any educational benefit from the program.Ph.D. (Doctor of Philosophy