Estimating the Overall Effects of Statewide Early Literacy Policies: Evidence from Michigan

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

Abstract

Over the past two decades, more than 40 U.S. states have implemented early literacy reforms aimed at ensuring students achieve reading proficiency by the end of third grade. While numerous studies have examined the local average treatment effects of third-grade retention mandates associated with these policies, state-level early literacy reforms typically include a broader set of supports and interventions that impact a wider range of students and teachers across grades K–3. Retention is often positioned as a last resort, implemented only after other policy tools—such as improved instruction and early identification of reading difficulties—have been exhausted. This paper investigates the broader impacts of comprehensive early literacy initiatives.​

Focusing on the implementation timeline of Michigan's Read by Grade Three (RBG3) Law, we employ a synthetic difference-in-differences method to address challenges arising from the lack of a suitable control group, given the statewide and simultaneous rollout of such policies. Using Civil Rights Data Collection and SEDA data, we assess the overall effects of Michigan’s early literacy policy on student achievement and grade retention, relative to other states without comparable policies. To contextualize Michigan’s results within the broader context of early literacy reforms nationally, we estimate equivalent models for two other states that enacted similar policies.​

 Our analysis reveals that Michigan’s early literacy policy positively influenced average reading test scores for all students in the state prior to the COVID-19 pandemic. This improvement was most pronounced in third grade, with an increase of 8.7% of a standard deviation in SEDA cohort-standardized scores, with higher scores for cohorts exposed to the policy longer. However, this effect diminished in higher grades, with gains of 5.3% in fourth grade and 2.7% in fifth grade. Furthermore, we do not observe any spillover effects of Michigan’s early literacy policy on math scores, shedding light on the potential complementarity between literacy policy and math outcomes.

Michigan's improvements in reading scores surpassed those of South Carolina’s Read to Succeed Act but were less pronounced than the gains achieved under Mississippi’s Literacy-Based Promotion Act. While Mississippi experienced a larger effect on third-grade retention compared to the negligible effects in Michigan and South Carolina, the magnitude and lack of persistence of this effect suggest that retention alone does not fully account for the differences in the estimated effects of these policies. Further investigation into the three states’ implementation strategies revealed several key factors, including the pace of policy rollout, length of exposure, specific policy funding, and complementary policies, which appear to contribute significantly to the variation in policy outcomes.​

Additionally, baseline literacy levels prior to policy implementation may explain some of the heterogeneity in policy effectiveness across states. We find a similar relationship across school districts within Michigan, where districts with the highest baseline achievement experienced the smallest gains in reading scores as a result of the policy.  Leveraging survey data on educators’ perceptions of implementation and granular financial information on district expenditures related to policy funds, we further assess how differences in implementation correlate with variability in effects.

Author

• Josep M Nadal-Fernandez, Tulane University
  Presenting Author