Michigan has spent a decade trying to improve how children learn to read. The latest results suggest there is still a long way to go. This spring, just 38.2% of Michigan third graders scored proficient or better in English language arts on the M-STEP. That comes after years of new laws, literacy coaches, teacher training and changes in how schools identify and support struggling readers.
I taught through much of this period, and I remember the attention that accompanied new education initiatives from Lansing. What was harder to see was what changed the next morning inside an individual classroom. That leaves an uncomfortable question: How long should it take for education reform to actually reach students?
Why are Michigan's reading results so troubling?
The problem is bigger than one disappointing test score. On the 2024 National Assessment of Educational Progress, commonly called the Nation's Report Card, Michigan fourth graders averaged 209 in reading, compared with 214 nationally. Michigan scored significantly below 31 states and jurisdictions and above only two.
Michigan fourth graders averaged 219 on NAEP in 2003. By 2019, before Covid disrupted schools, the score was 218. In 2024, it fell to 209.
There is no single accepted explanation for Michigan’s struggles. Poverty, achievement gaps, absenteeism and pandemic disruption all matter. So what happened to all the reforms we have already tried?
Why haven't ten years of reform fixed that?
One problem with describing Michigan's “decade of reading reform” is that it sounds as though the state chose an approach in 2016, implemented it and then waited ten years for results. That isn't what happened.
The 2016 Read by Grade Three law emphasized early screening, intervention, literacy coaching and, most controversially, retaining some third graders who were far behind in reading. The retention requirement was later repealed.
Michigan changed direction again with new literacy and dyslexia laws in 2024, which revised and expanded screening, intervention and professional-learning requirements. Some of those provisions do not fully take effect until the 2027-28 school year.
Then came the 2026 Science of Reading, which requires K-5 educators involved in reading instruction to complete state-funded professional learning in the Science of Reading. The deadline is the beginning of the 2031-32 school year.
Michigan has not spent a decade implementing a literacy policy, it has spent a decade reforming literacy policy.
That distinction matters when we look at test scores. It would be unfair to judge a reform that hasn't been fully implemented by test scores produced before implementation is complete. At the same time, after a decade of reform, why does Michigan remain years away from fully implementing its latest approach.
Why is implementation so difficult?
Changing a law is easier than changing what happens between a teacher and a student. The state can require assessments, approve curricula, mandate professional development and collect test scores. Those things are important, but having a teacher complete training isn't the same as changing instruction in the classroom.
A recent University of Michigan study of the state's early literacy coaching initiative illustrates both the promise and the difficulty. Researchers Tanya Wright and Lori Bruner found that coaching helped teachers improve their use of research-aligned literacy practices. They also found that high-quality implementation was possible on a large scale. But local conditions affected how consistently those practices were adopted. Reform can reach the classroom, but doing it consistently across thousands of classrooms is considerably harder.
My own experience was with older students. I taught high school history, not reading, yet I regularly encountered students who could read the words on a page but struggled to understand what they meant. Sometimes I stopped teaching history content and instead worked on reading comprehension: breaking down a difficult passage, developing background knowledge, working through unfamiliar vocabulary or helping students identify an author's argument.
No one told me not to do that, or suggested comprehension wasn't important. But no one required me to do it, either. I was accountable for teaching history content. Time spent helping students become better readers was time not spent getting through that content.
That distinction shouldn't be mistaken for a criticism of the Science of Reading. Michigan defines the approach as including phonemic awareness, phonics, fluency, vocabulary and comprehension. My concern is about something broader. In any large system, the things that can be readily required, tracked and tested have an obvious advantage over something so difficult to observe as whether an individual student is becoming a better reader of complicated material. My classroom experience makes me wonder how much that shapes what ultimately receives attention.
Why does the delay matter so much now?
If changing instruction takes years, students don't get to wait. A third grader struggling with reading today will be in middle school before some of Michigan's newest literacy requirements are fully implemented. And while that child moves through school, the world in which literacy matters is changing too.
Today's students face an information environment overflowing with online material, misinformation and increasingly AI-generated content. Being able to read the words is not enough. They must understand difficult information, connect it with what they already know and decide whether what they are reading deserves to be believed.
Artificial intelligence complicates that task further. AI can be enormously useful to a struggling reader. It can define unfamiliar words, provide background information, simplify a difficult passage or explain an argument. It can also summarize the passage so the student doesn't have to wrestle with it at all.
A 2026 randomized study of 405 14- and 15-year-old students in England found that students generally preferred using large language models and considered them helpful, but students who took notes, either alone or while using AI, showed better comprehension and retention than students who relied on AI alone. That captures the paradox: AI may become one of the most powerful tools ever available for helping struggling readers, while also becoming one of the easiest ways to avoid some of the intellectual work through which stronger reading develops.
The skills students need are changing accordingly. The OECD's new framework for AI literacy emphasizes not merely using AI but critically evaluating its output. Michigan is trying to improve literacy while the meaning and demands of literacy are themselves changing.
Michigan's newest literacy reforms deserve enough time to determine whether they work. Its poor reading results deserve urgency. Those ideas aren't contradictory.
Education reform moves at the speed of institutions. Childhood moves at the speed of childhood. Technology now moves faster than both.
Michigan's challenge isn't simply finding the right literacy policy. It is narrowing the distance between good policy and classroom practice fast enough to matter to the children who are there now, while preparing them for a world that won't wait for schools to catch up.
