Longitudinal Impacts of Oversubscribed Boston Pre-K Programs: Evidence from Students’ Elementary, Middle and High School Years

• In Event: The Long-Term Effects of Preschool Interventions

Abstract

Children who attend prekindergarten programs show clear short-term gains, but researchers continue to debate how long these benefits last and under what conditions (Cascio, 2021). Boston Public School’s (BPS) high-quality, school-based Pre-K offers a strong test of lasting impacts, with an unusually strong counterfactual in which almost all children in the control group attended other preschool programs. The program implemented evidence-based language, literacy, and math curricula, provided curriculum-specific coaching for teachers, and treated Pre-K teachers with parity with K-12 counterparts. Leveraging oversubscribed first-choice lotteries, our preregistered exploratory study examines: What is the impact of Boston Pre-K on children’s educational trajectories, school progress/engagement, and academic achievement in late elementary, middle school, and high school?

We analyzed data from four cohorts of students who applied to BPS Pre-K between 2007-2008 and 2010-2011 (N=3,092; 52% female, 8.2% Asian, 22.8% Black, 39.2% Latino, 25.9% White). Students who won the lottery for their first-choice school constitute the treatment group; children who lost their first-choice lottery comprise the control group. We linked BPS and state administrative records to track this sample from Pre-K-12th grade. We leveraged first-choice lotteries for BPS Pre-K seats to estimate the intent-to-treat effects of winning a first-choice lottery (ITT; Abdulkadiroğlu et al., 2011; Bloom & Unterman, 2014) and complier average causal effects of enrolling in Boston Pre-K (CACE; Gennetian et al., 2005). We estimated ITT impacts using a regression with lottery fixed effects, student covariates, and standard errors clustered by students’ Pre-K school. For the CACE, we used 2SLS with the first-choice lottery offer as the instrumental variable (compliance rate≈29 percentage points).

There were substantial treatment impacts on children’s educational trajectories (BPS enrollment and persistence, charter school enrollment and persistence, exam school application and enrollment), in K-8 but not high school. Results show evidence of positive impacts on children’s math achievement for treatment compliers, particularly on enrolling and passing Algebra I by end of 8th grade (22.6pp difference, p<0.01), 7th and 10th grade math test scores (0.24SD and 0.31SD, respectively, p<0.05), and SAT math scores (40 points, p<.05). Treatment compliers were less likely to be suspended in 2nd and 7th grades (-4 and –8, p<.05) and more likely to take and pass an AP exam in high school (14 and 17pp, p<.05). There were no impacts on students’ retention, special education placement, reading, nor on taking the SAT, while findings for attendance were not robust. Generalizability analyses suggest larger benefits for the full sample of enrollees. Mechanisms analyses suggest children’s schooling environments after kindergarten and the skill boost Pre-K provided are both channels for our detected effects.

Our findings show that the effects of high-quality Pre-K programs can last into the K-12 years. Because nearly all children in the control group attended other preschool programs, our findings also provide suggestive evidence on the important question of what kind of preschool can produce a lasting boost. Finally, our mechanisms findings underscore the value of building “sustaining environments” (Bailey et al., 2017) that bridge Pre-K and later grades to allow early skill advantages to compound.

Author

• Rebecca M Unterman, MDRC
  Presenting Author