Grouping Students by Ability Transformed Our School. Could It Do the Same for Yours?

A Story of Educational Progress from an elementary school in Rockford, Illinois.

When I wrote about Ellis Elementary's shift to grouping students by reading ability instead of grade level for The 74, I had no idea how far that story would travel. I didn’t expect lengthy “comment section” conversations or thoughtful questions landing in my Inbox from educators, administrators, and policymakers across the country. It was clear that people were curious and hungry for a blueprint for how we did it.

This follow-up piece is just that: the nuts and bolts of how we made this work at Ellis, and how it might work at your school, as well. It’s for everyone who has asked about the logistics, policy flexibility, cultural shift — and the inherent challenges behind our success. Because when our school decided to stop grouping students by grade level for reading instruction and instead group them by ability, it wasn’t because we had all the answers.

We had run out of excuses. Our reading and math scores were in the bottom 5% of schools in Illinois. The old ways weren’t working. Our students were behind, and COVID only widened the gap. If we wanted different results, we had to do something different. That decision led to dramatic improvements in our students’ reading achievement.

My article generated as much interest as it did questions. Educators and policymakers wanted to know how we did it. What did implementation really look like? What were the pitfalls? How did we get buy-in? Why not replicate it across other subjects? So here’s the follow-up: a deeper dive into how we made subject-specific ability grouping work for reading, what we’re learning as we apply those lessons to math, and what advice I’d give to any school that knows it needs to change but isn’t sure where to start.

---

A Day in the Life at Ellis

We built our school day on the belief that the right instruction, at the right level, delivered with consistency, makes all the difference. And that curriculum matters. We use Reading Horizons for phonics instruction and CKLA for knowledge-building and comprehension.

Here’s what a typical instructional day looks like for students at Ellis:

• Tier 1 Phonics Instruction: 45 minutes

• ELA Small Groups: 60 minutes

• Tier 1 Knowledge-Building ELA: 90 minutes

• Tier 1 Math: 45 minutes

• Flex Math: 30 minutes

Students travel to ability-based groups for Reading Horizons instruction (foundational literacy skills, including phonics and phonological awareness). These groups are not tied to their homeroom, and the teachers rotate accordingly. For the most part, students stay with their homeroom teacher for the rest of the day and receive targeted support in smaller groups (as needed).

Making Ability Grouping Work for Reading

With any ability grouping, the first step is identifying each student's ability level. We started with a student-facing diagnostic to gauge each student's decoding and encoding abilities.

Then came the hard part: organizing the groups. We aimed to avoid moving students more than one grade level below their actual grade for instruction (if possible), both for academic and social-emotional reasons. We stacked each grade level with a lower-skill group so students could stay in a familiar environment. We intentionally kept the lowest-performing groups small to ensure they’d receive more individualized attention. And we embedded the belief that these aren’t "your kids" or "my kids" — they’re all our students.

Routines and expectations were critical. Every classroom follows the same procedures for entering and leaving the room, getting materials, and sharing out, so students know what to expect regardless of which teacher they're with. We made sure every teacher understood the urgency and supported the shift. As I often say, you’re only as strong as your weakest link.

Initially, we didn’t include kindergarten in the ability grouping model. That changed quickly when we realized the value of accelerating our youngest students who were ready for more of a challenge.

So What About Math?

After a few years, we began to see significant momentum and growth in reading, so we shifted our attention to math. Our math data confirmed what we already suspected: most students were at least one grade level behind.

But math posed new challenges. There wasn’t a neatly packaged, small-group curriculum similar to popular choices currently available for ELA; we didn’t have a ready-made scope and sequence for differentiated math instruction.

So we built one.

Within each grade level, using our tier 1 math diagnostic data, we divided students into three groups: two grades below, one grade below, and on grade level. We assigned each group to a specific teacher and provided whole-group instruction on prerequisite skills that aligned with the Tier 1 math lesson they would encounter later in the day. Our goal for both ELA and math is to front-load prerequisite skills with targeted intervention before whole-class Tier 1 instruction, not after. It’s a proactive model, not reactive remediation.

Next year, we plan to introduce a new component called Fluency in Five, inspired by Brian Poncy’s M.I.N.D. resources, which focuses exclusively on math fact fluency.

Although we’ve seen small gains on math state testing, we haven’t yet seen the significant gains we’ve had in reading; however, we’re laying the foundation. In many ways, we’re still building the plane while flying it. We believe that this model, which pre-skills students before Tier 1 instruction, can be effective across various subjects. We just haven’t had our “Sold a Story” moment for math — yet.

Addressing Concerns: Stigma, Staffing, and Student Belonging

I often get this question: What does this do to student self-esteem, especially for third-graders grouped with first-grade-level readers?

Honestly, it hasn’t been the issue some feared. Part of that is because of the way we communicate with students. We don’t make the groupings feel punitive. Part of it is the consistency of instruction across classrooms. But a big part of it is that for many of our kids, this is the first time school has made sense.

And since we’ve been grouping this way since 2019, for many students, this is all they’ve ever known. It’s the norm, and it works for them — so they don’t question it.

In truth, ability-based reading instruction with Reading Horizons is often our students' favorite part of the day. When students feel successful, the need to act out or save face evaporates. We have significant behavior supports at Ellis due to our students’ diverse needs, but we haven’t found that ability grouping creates new discipline problems. If anything, it reduces them.

Yes, there is a staffing demand, but not due to behavioral issues. It’s about group size. The more teachers we have, the smaller and more targeted our groups can be. That requires coordination and support, which brings me to policy.

Ellis is part of a large district with over 30,000 students. The district chooses our curriculum, but we have some autonomy over how we deliver instruction. That policy created the opportunity we needed.

In districts with strict policies, begin with a small, strategic pilot. First, build support from your school leadership by highlighting successful implementations in schools with similar demographics and challenges. Then, identify champions of change within your building and launch the initiative in one grade — or even a single classroom. Establish a clear vision and collect data to track student outcomes. Once you have results, present your findings to district leaders or the school board to advocate for broader adoption.

In our district, with early success, other schools began to take notice. Several have now adopted "walk to learn" models like ours (i.e., students moving to a different classroom to receive ability-based instruction). And our district leadership has been supportive. They’ve let us lead and continue to allow us to go where the data takes us.

What Made This Work

I always get asked about resistance. From school leaders. From teachers. From parents.

And here’s the truth:

We were in the bottom 5% in the state. What did we have to lose?

We built teacher buy-in by building teacher confidence. We showed them data. We provided them with evidence-based curricula and the necessary training to use them. We created consistency. That led to more confident instruction, more consistent student experiences, and results we could be proud of.

If you’re on the fence about trying something like this, here’s my advice: change is scary, but necessary, especially when your most vulnerable learners are at stake.

• You don’t need the perfect packaged solution. We didn’t have one for math, so we made our own. You just need the will to stop doing what isn’t working, the flexibility to find what does, and the humility to keep iterating.

• Identify the areas where your policies permit freedom, even in a large district like Rockford. Start there. Don’t wait for permission (within reason, of course).

• Continue evolving. We’re not done. Our math grouping model is still evolving. We’re exploring ways to enhance fluency and provide better prerequisite support. We’re also opening up our data for deeper analysis, including how different ability groups are growing over time.

But if there’s one thing I’ve learned, it’s this: once you stop asking "What if it doesn’t work?" and start asking "What if it does?", the path forward becomes a lot clearer.

---

Jessica Berg has served as the Instructional Coach at Ellis Elementary School in Rockford, IL for 7 years. She is also trained in both LETRS and Orton-Gillingham (OG). Her passion for education and the Science of Reading was sparked by her son’s diagnosis of dyslexia at age 14.

She also writes on Substack!

And to support more content like this from the Center for Educational Progress…