Strong early literacy skills are crucial for children’s school and life success. However, the latest results from the National Assessment of Educational Progress reveal that only about a third of American fourth graders read at a proficient level, with children from historically marginalized backgrounds trailing behind their peers. As educators work to improve reading outcomes, requiring students to repeat a grade—known as retention—has become a common, though controversial, component of state policy reforms, with mixed evidence on student outcomes.

As states adopt “science of reading” reforms and districts struggle to undo pandemic-related learning losses, grade retention policies have become more common. As of 2025, 14 states covering 34 percent of public school third graders require students to pass a reading test to be promoted to fourth grade. Additionally, four other states, including Maryland and West Virginia, have also passed third-grade retention laws that will soon take effect. Compared with older retention laws, modern ones generally make greater use of “good cause” exemptions that allow students to be promoted despite being flagged for retention. Modern policies also require schools to provide flagged students with more reading support, whether or not they are ultimately retained.

Our research examines the effects of one of these recent reforms: Michigan’s Public Act 306 of 2016, also known as the Read by Grade Three Law. We used statewide administrative data to study the effects of the law on the first children subject to retention: third graders in the 2020–2021 and 2021–2022 school years. According to the law, third graders who scored below a predetermined cutoff on the state reading test in the spring were flagged as eligible for retention. Their families were informed over the summer via a letter from the state. Our analysis compares children who scored slightly below the cutoff score with similar children who scored slightly above it. Differences in outcomes between these groups reflect the effects of being flagged for retention, including all reading interventions and behavioral responses that followed.

Our findings reveal that being flagged increased the likelihood of being retained by only 3.4 percentage points. Indeed, only 6.8 percent of flagged third graders were retained. Nevertheless, being flagged increased students’ reading test scores in the following school year by 0.045 standard deviations, on average—a small but educationally meaningful effect. Our findings are similar for most demographic groups, but we did find some evidence that children from historically advantaged groups, including native English speakers and families with higher incomes, may have benefited more from the policy—especially because they became more likely to have an IEP (individualized education program) in the next school year.

Policymakers and educators might also want to know the effect of actually retaining students rather than simply flagging them. Yet, applying conventional analytical methods would suggest implausibly large gains from retention: a 1.3 standard deviation increase in reading scores. For comparison, typical growth in reading between third and fourth grade is only about 0.36 standard deviations. This finding led us to examine a previously overlooked feature of most modern state retention laws that would explain our implausibly large estimate.

Specifically, most retention laws (including Michigan’s) require that all flagged students receive reading interventions in the following year, regardless of whether they are retained. These interventions, along with teachers’ and families’ responses, violate the standard analytical assumption that flagging only affects student outcomes through retention. Indeed, principal survey data from Michigan reveal that schools often provided extra help—such as summer reading programs and high-dosage tutoring—to students who scored just below the retention cutoff even if they were not retained. Moreover, our research finds positive effects from flagging even in districts that did not retain any students, further demonstrating that flagging students improved their literacy skills through channels other than just retention.

Our findings provide several insights for policymakers. First, because Michigan’s reading reform closely resembles those in other states, some prior studies likely face the same analytical issues we discovered, so their effect estimates may be misleading. Second, our research finds that flagging students for retention improved their reading test scores, though retention itself played a minor role. This finding highlights the role of other reading support services often included in state reading policies. These services receive less attention than retention, but they are particularly relevant in states (such as Maryland) that are designing policies with less emphasis on retention. Michigan eliminated the retention component of its Read by Grade Three law in the 2023–2024 school year but maintained the non-retention support services. Overall, our findings suggest that grade retention is a much less important component of state reading law reforms than previously assumed.

Note
This research brief is based on Jordan S. Berne et al., “The Impacts of Grade Retention Policy with Minimal Retention,” Annenberg Institute at Brown University EdWorkingPaper no. 25–1188, May 2025.