Examining Heterogeneity in the Impact of the North Carolina Pre-K Program Across Characteristics of Children and their Families

• In Event: Disentangling Heterogeneity in the Impact of North Carolina’s Public Pre-Kindergarten Program

Abstract

One robust conclusion from the evidence on scaled-up pre-k programs is that the long-term effects of experiencing pre-k are heterogeneous. This has piqued interest in examining explicitly the ways that impacts vary, with sources of heterogeneity involving both characteristics of the child and their environments. Consistent with the theory and empirical work on the differential and compensatory effects of early education, several characteristics of families and children might predict who will benefit most from pre-k.
Our study’s goal is to understand heterogeneity in the long-term impacts of the statewide pre-k North Carolina’s Pre-K Program (NC Pre-K) by examining whether and how pre-k effects differ across groups of children that vary in race, ethnicity, birthweight, sex, and mother’s education. These factors represent individual strengths, but also systemic environmental disadvantage and disenfranchisement.

We build upon the data and analytic strategy described in Dodge, Ladd, Muschkin and colleagues (2017; 2014; 2015), and use a difference-in-differences analytic approach that exploits exogenous variation across counties and years in the introduction and allocation of state funds to counties for the program to evaluate impact. We examine variation by child and family characteristics through four sets of population subgroup analyses to test for heterogeneity: Race and ethnicity (Black, Hispanic, White, and other racially identified students); sex (males, females); birthweight (low, regular/high); and mother’s education (>high school, high school+). We estimated our primary specification used to assess the main effects separately for each of the subgroups, and tested for statistical differences in the coefficient on NC Pre-K across groups by estimating models interacting NC Pre-K and the subgroup characteristic. Our sample is the population of children born in NC between 1987 and 2005 who later attended 5th grade in a public school in that state and could be matched by administrative record review (n = 1,215,488; 58% White, 29% Black, 7% Hispanic, 6% Multi-racial and Other race/ethnicity), which includes six additional cohorts not in prior studies. Our outcomes are academic achievement (math and reading scores; composite), and grade retention and special education status, measured in 5th grade.

Table 1 presents our results where each column restricts the analytic sample to each demographic subgroups. NC Pre-K funding had a positive and statistically significant effect on achievement for children who had normal birthweight, whose mothers did not have a high school diploma, were Black, Hispanic, males, and females. The effect for White children and children whose mothers had a diploma was positive and marginally significant. The effect was not significant for low birthweight children and children from an “other” race/ethnic group.

These estimates were all significantly different from one another. For special education and retention, there were no significant relations with NC Pre-K funding for any of the subgroups.
We further probed heterogeneity intersectionally, examining whether children who identify as members of multiple subgroups have varied impacts of NC Pre-K funding (Table 2).

Generally, we found that across these various combinations of groups, effects were still larger for Black and Hispanic children, and children with mothers with less education.

Authors

• Jade V Marcus Jenkins, University of California - Irvine
  Presenting Author

• Tyler Watts, Teachers College, Columbia University

• Kenneth A Dodge, Duke University

• Robert C. Carr, University of North Carolina at Chapel Hill

• Maria Sauval, University of California - Irvine