"The Impact of the Chicago School Readiness Project (CSRP) on Adolescent Executive Functioning and Behavior"
AERA 2018 Annual Meeting: “The Dreams, Possibilities, and Necessity of Public Education”
April 13-17, 2018
Title: The Impact of the Chicago School Readiness Project (CSRP) on Adolescent Executive Functioning and Behavior
Authors: Tyler Watts (PhD Alumnus), Jill Gandhi, C. Cybele Raver
Researchers and policy-makers alike have paid a great deal of interest to early childhood education (ECE) programs. Recently, much of this attention has focused on the long-run effects of ECE programs, as some hallmark studies showed long-run effects into adulthood (e.g., Perry Preschool, Abecedarian). More recent investments in early childhood programs have yielded positive short-run benefits followed by long-run “fadeout,” i.e. no evidence of impact, several years later (see Bailey et al., 2016).
The current study evaluated the long-run effects of the Chicago School Readiness Program (CSRP), which was tested with 602 preschoolers enrolled in 18 Head Start centers in Chicago (see symposia description for intervention details). Raver and colleagues (2011) reported large and positive effects on students’ levels of school readiness skills (e.g. measures of executive function and pre-academic achievement) at the end of preschool, but effects had largely faded by elementary school. In the following analyses, we report impacts of the CSRP preschool program on 461 adolescents’ neurocognitive and behavioral development, measured ten years after initial assignment.
We began by evaluating baseline equivalence and patterns of attrition. As Table 1.1 reflects, about 20% of the sample had left the study by the ten year follow-up, but this rate was equivalent between both groups. We evaluated baseline equivalence across several observable characteristics. We found no indication of baseline imbalance across most characteristics. We present treatment impacts from OLS models for measures of CSRP-enrolled adolescents’ executive functioning and behavioral problems. The executive functioning skills of working memory and inhibitory control were measured using the Hearts and Flowers (H&F) task, and we examined both accuracy and reaction time on “mixed” trials (i.e., the most difficult trials). Internalizing and externalizing behavior problems were measured using a youth self-report.
For all outcomes assessed, we started with bivariate treatment impact models. We then moved to models that controlled for blocking group, as random assignment was conducted through a blocking procedure that paired sites on several characteristics. In columns 3 through 5, we added baseline characteristics to adjust for any bias due to imbalance at random assignment. We found positive effects of assignment to CSRP on H&F accuracy, with the fully controlled estimate having a positive effect of about 1/10th of a SD (b= 0.13, p < 0.10). This estimate was fairly consistent across controlled and uncontrolled models, indicating that any imbalance at baseline had little effect on observed impacts. We found no effects of the treatment on H&F reaction time, and limited evidence of treatment impacts on behavioral problems.
Our results indicate that the CSRP had a long-run effect on a key measure of students’ executive functioning ten years after enrollment in the preschool intervention. Future analyses will focus on possible sources of treatment heterogeneity (e.g., level of poverty at baseline) and we hope to extend our analyses to other outcomes (e.g., academic achievement).