Does Early Childhood Education Help to Improve High School Outcomes? Results from Tulsa

• In Event: What Works Best and For Whom: Treatment Heterogeneity in Long-Term Effects of Early-Care and Education

Abstract

In recent years, early childhood education (ECE) scholars have turned attention to long-term effects of state-funded and federally-funded preschool programs. Many studies show persistent positive effects of high-quality programs, but the Head Start Impact Study and Tennessee Voluntary Pre-K Study have produced contrary findings (Lipsey, Farran, & Hofer, 2015). Early studies, like the Perry Preschool Study and Abecedarian Projects, found positive effects many years after the intervention (Campbell, Ramey, Pungello, Sparling, & Miller-Johnson, 2002; Heckman, 2006). Most contemporary programs, however, are too young to be compared to these iconic programs. The Tulsa pre-K study is a rare exception—now in a position to analyze high school outcomes, comparing students who completed the Tulsa Public Schools pre-K program and the CAP of Tulsa County Head Start program with a control group of students who did not attend pre-K or Head Start.
Students in our sample started kindergarten in Tulsa Public Schools in the fall of 2006; about 40% of them attended pre-K and about 11% attended Head Start. In 2006, we obtained data on a range of demographic characteristics, such as gender, race/ethnicity, and special needs status. Parents also completed a survey that covered maternal, marital, and education status, presence of father and Internet at home, language spoken, and books at home. State data include achievement test scores (SAT or ACT), attendance, and grade retention information. School district data, from the Tulsa metro area, include courses taken and grades received. We used baseline covariates to generate propensity scores and included several of them as covariates in regression models.
We employed multiple imputation, which estimated missing values on covariates using chained equations, filling in missing data as estimates from complete data in an iterative process (Little & Rubin, 2014). We used propensity-score weighted multiple regression to estimate the effects of ECE during the 2017-18 academic year, when most of our students were high school juniors. The current analyses examined enrollment in Tulsa’s school-based pre-K program and Head Start as predictors of students’ total number of absences, whether a student was chronically absent, ACT/SAT math and English Language Arts (ELA) scores, whether a student was enrolled in an AP or IB course, and academic GPA.
Tables 1 and 2 present findings for the effects of TPS pre-K and Head Start participation, respectively, on students’ academic outcomes during the 2017-18 academic year. TPS pre-K participation was associated with fewer days absent, increased odds of being enrolled in an AP or IB course, lower odds of failing a course, and lower odds of being retained in grade. There were no significant associations for Head Start. Subgroup models suggest differential effects by race, ethnicity, and gender. Specifically, children of color are more likely to experience long-term benefits by the end of their junior year of high school than other children.
The analyses provide evidence that pre-K participation may contribute to positive high school outcomes much later. This work, along with other contemporary studies, will continue to illuminate whether and how ECE programs contribute to educational achievement.

Authors

• Sara Amadon, Child Trends -Maryland
  Presenting Author

• William T. Gormley, Georgetown University
  Non-Presenting Author

• Katelyn Romm, George Washington University
  Non-Presenting Author

• Katherine Magnuson, University of Wisconsin - Madison
  Non-Presenting Author

• Amy Claessens, University of Chicago
  Non-Presenting Author