A Mississippi Miracle? The Effect of Mississippi’s Science-of-Reading Reforms on Elementary Reading Skills

• In Event: Science of Reading Legislation's Effects and Their Mechanisms: Studies of Legislation from Three States

Abstract

Purpose
In 2013, Mississippi passed the nation’s first science-of-reading-based legislation: the Literacy-Based Promotions Act (LBPA). Over six years, state NAEP scores rose 10 points, inspiring 40 states to pass similar legislation. Despite the LBPA’s influence, no causal study has investigated its effect or mechanisms behind it. This study targets this gap through a difference-in-difference (DiD) analysis isolating the effect of the “Literacy Support” (LS) intervention designed to improve teacher instruction through professional development and coaching.

Perspectives
To understand whether and how the LBPA improved student outcomes, we ground this study in features of the legislation, including:
• Prohibiting social promotion to 4th grade unless students exceeded a threshold on a 3rd grade reading test or had a “good cause exemption” (e.g., SWD).
• Improving reading instruction and intervention through LETRS PD, use of curricula aligned with the science of reading, early and frequent screening to identify struggling readers and provide intervention.
• Lowest-performing schools received LS intervention including a half-time coach and mandating K-3 teachers attend LETRS PD (Moats & Sedita, 2004).
Although it’s reasonable to assume that reformed instruction drove reading improvements, other possible mechanisms include Mississippi’s 2013 preschool subsidy and the end of social promotion. Data displayed in Figure 1 drives our hypothesis regarding social promotion.
Methods
A two-way fixed-effects model allowed us to estimate potential causal outcomes. Because LS rolled out gradually (Figure 2), we asked whether early-LS-schools experienced earlier or larger reading improvements using the staggered DiD model:

Table 1 displays details regarding data sources. We obtained data from Mississippi Department of Education website, via public information requests, and federal Common Core of Data.

Results
Figure 3 shows trends in statewide average kindergarten literacy scores in fall and spring. Between 2014-2015 and 2015-2016, there was a large jump in spring test scores—an improvement that persisted through 2018-2019. Because fall test scores show similar entry skill levels, this trend suggests improvement in instructional quality during kindergarten.
Figure 4 displays fall and spring trends by schools’ LS start years. There’s no clear correlation between LS start year and spring scores. Across Mississippi, kindergartners graduating after 2015-16 achieved higher reading, regardless of LS. However, the 2015-16 spring score jump was greater for earlier LS cohorts than for those joining after 2017-18 or receiving none, suggesting statewide reading reform (e.g., LETRS PD) had broad impact. LS contributed but likely didn’t drive improvements.
Our next question became whether teachers outside of LS schools received LETRS PD, and if broad participation explains effects. We recently received lists of PD participants and will model this variable; results will be available for the conference.

Significance
As Mississippi inspired many states’ reading legislation, information regarding what caused their improved scores is likely invaluable to states trying to replicate results. This study accordingly presents a realization of the conference theme of better implementing policy. Our results suggest mandated PD, coaching, retention do not fully explain improved scores, but the package of reforms—and potentially LETRS PD taken by teachers outside LS—has a positive effect in kindergarten.

Author

• Paul T. Von Hippel, University of Texas at Austin