The effects of scaling high quality curriculum

• In Event: Effectiveness of Early Literacy Policies: Evidence from Four States

Abstract

Increasing the quality of curricular materials has long been thought of as a promising strategy for improving learning, but several barriers prevent the realization of this promise. Teachers report spending up to seven hours per week—250 hours per year—sourcing or creating materials, but research shows homegrown materials are often lower quality, unaligned to standards and lacking on-grade level learning opportunities (The Opportunity Myth, 2018). They also create inconsistent learning experiences across classrooms. Even when districts opt to align instruction by providing prepared materials, the market for curriculum is large and it can be difficult to discern quality (Blazar et al., 2020; Polikoff, 2015). This suggests promise for policies which improve the quality and consistency of curricular materials across classes by helping districts navigate these complex markets. 

In this paper, I study a 2019 Rhode Island law which mandated the use of state approved math and ELA curriculum in all K-12 classrooms by 2025. The state department of education provided a list of approved curriculum products meeting quality thresholds and aligned to state standards. To meet state standards, ELA curriculum had to be rated as aligned with the science of reading. Moreover, all ELA teachers were required to demonstrate proficiency in the science of reading (SoR) by attending 50 hours worth of professional development focused on SoR principles. This constrained districts’ choice, but the approved list included at least 30 products which provided districts flexibility to select a product suited to their context. I leverage the staggered adoption of new curricular materials across school-grade-subject instructional teams to estimate average treatment effects of the policy on student learning outcomes.

Data. I use student-level administrative data from the Rhode Island Department of Education spanning the 2017-18 to 2023-24 academic years in grades 3-8, 9, and 11. My primary outcome is student achievement in math and ELA on state achievement tests in grades 3-8 and PSAT and SAT exams in grades 9 and 11, respectively. 

Methods. I estimate the average treatment effect of choosing materials from the state-approved list and adopting them on student achievement through a generalized difference-in-differences approach using two-way fixed effects and event-study models (Baker et al., 2025; Wing et al., 2024). I also conduct exploratory heterogeneity analyses to test whether features like teacher experience, teacher productivity, and the curricular materials in use prior to adoption moderate treatment effects. 

Results. I find the policy significantly improved math achievement by 0.07 SDs but had a relatively precise null on ELA achievement. Math effects were driven by significant gains of 0.09 SDs among instructional teams with above average predicted student achievement at baseline. I find these instructional teams benefitted immediately from the policy with effects persisting over time. I find suggestive evidence of a required familiarization period in similarly above-average ELA instructional teams, however effects are statistically insignificant. I find no effect on average among instructional teams with below average predicted student achievement at baseline, but significant gains of 0.05 SDs among teams switching from other curriculum products rather than locally developed materials. 

Author

• Alexander J Bolves, Harvard Graduate School of Education
  Presenting Author