The Impacts of Grade Retention Policy With Minimal Retention

Jordan S. Berne, Brian A. Jacob, Christina Weiland, and Katharine O. Strunk

This study evaluates Michigan’s third-grade reading policy, which flags students with low reading scores for potential retention. Although few students were actually retained, being flagged led to a small but statistically significant gain in reading achievement, about 0.045 standard deviations (SD), in the following year. The positive effects appear driven not by retention itself, but by increased support from schools and more engagement from families once students were flagged.

The study underscores the importance of moving beyond the retention/no-retention debate. Instead, policymakers should focus on how to identify struggling students early and deliver the right supports.

The Policy

In response to national concerns about early literacy, 18 states as of 2023-24 required test-based grade retention in third grade. Michigan implemented its "Read by Grade Three" law in 2016, mandating retention for students scoring below a specific reading threshold, but also offering “good cause” exemptions and mandating extra support for struggling students. In 2023, Michigan dropped the retention portion of the law.

Sample and Methods

This study used Michigan's administrative data and applied a statistical technique called regression discontinuity. This technique compares students who score just below the cutoff (and are flagged) with those who score just above it (and are not flagged). These two groups of students are nearly identical in every way except for whether or not they were flagged for retention. Because students on either side of the cutoff are so similar, the only meaningful difference between them is the policy's treatment: being flagged for retention. This setup mimics a randomized experiment, where assignment to the "treatment" (being flagged) is essentially random for students near the cutoff.

Key Takeaways for Policymakers

1. Being flagged for retention, more than retention itself, drove improvements in literacy.
   Third-grade students flagged for retention due to low reading scores scored higher in ELA in the following year compared to similar non-flagged students. Although Michigan’s policy included a retention mandate, flagged students improved even when they were not retained, suggesting most of the academic benefits came from being flagged as at-risk and the increased supports that came with being flagged.
2. Policymakers should consider decoupling supports from retention. 
   Since being flagged for retention led to higher reading scores even when students were not retained, states may be able to achieve similar or better academic outcomes by offering supplemental services (e.g., tutoring, summer programs, extra instructional time) without holding students back. This approach could reduce large-scale retention, which is often unpopular and difficult to implement at scale.
3. Policymakers and researchers should invest in tracking what services are provided, to whom, and with what intensity.
   This would allow us to move beyond simple “retained vs. promoted” comparisons and instead evaluate what actually works to improve student outcomes. Currently, many evaluations of reading retention laws treat the policy as one large bundle of interventions, combining retention and other supports. By collecting student-level data on the support individual students receive, states could begin to identify which supports drive academic gains and for which types of students, and ultimately build more effective support systems.

Key Findings

1. Higher ELA Achievement: On average, students flagged for retention saw a 0.045 SD gain in English Language Arts (ELA) scores in the following school year (4th grade).
   This tells us that the policy of flagging students triggered real changes that caused improved literacy outcomes, compared to what would have happened if those same students had not been flagged.
   • Positive effects were found for White and Hispanic students, but not for Black or Asian students.
   • Students without an IEP experienced a small gain in achievement (0.064 SD). In contrast, students with an IEP saw no benefit as a result of being flagged for retention.
2. No Clear Effects on Math, Attendance, or IEP: Being flagged had no significant effect on attendance or the likelihood of receiving an Individualized Education Program (IEP) in the following year. Impacts on math scores are inconclusive but may have been positive.
   • Non-economically disadvantaged students were more likely to receive an IEP the following year when flagged; economically disadvantaged students were not.
   • English learners (ELs) flagged for retention were less likely to be assigned an IEP, possibly because they received additional EL services instead.
3. Support Services, Rather Than Retention, Drive Impact: Positive effects were observed even in schools that did not retain any students.
   • This suggests extra literacy supports, not grade repetition, drove the gains.

The Mechanisms Likely Driving Higher Reading Scores

Being flagged doesn’t always lead to actual retention. In fact, most flagged students were promoted but still received additional supports like summer reading programs or extra help in class. That means this impact likely reflects a bundle of responses: parent engagement, extra school support, and increased teacher attention, not just the threat of retention. 

• Parent Involvement: Flagging triggered a state-mandated letter to parents, which may have prompted them to seek extra support for their child, inside or outside of school, regardless of whether the student was retained.
• School-Provided Supports: State law required that students flagged for retention receive additional literacy support, even if promoted. Common supports included:
  • Summer reading programs
  • Extra instructional time in literacy
  • Home reading support with families
• Potential Teacher Response: The retention flag may also influence teacher behavior, positively by prompting support or negatively by lowering expectations, though the study lacks direct evidence on teacher responses.

Full Working Paper

This report is based off the EdWorkingPaper “The Impacts of Grade Retention Policy With Minimal Retention” published in May 2025. The full research paper can be found here: https://edworkingpapers.com/ai25-1188