Student Grouping Strategies: What my dissertation taught me and how it changed my teaching

A teacher-friendly look at what national data suggest about grouping students in middle school mathematics (and what you can do differently tomorrow)

If you've spent any time in a math classroom lately, you've probably heard the message loud and clear: students should be collaborating. Administrators look for it during walkthroughs, curriculum writers build lessons around it and professional learning communities spend hours discussing how to improve it. However, collaboration and grouping aren't actually the same thing. Collaboration is what students do. Grouping is how we, as teachers, organize students to make that collaboration possible. It's a distinction that matters because while we spend a lot of time discussing collaborative learning, we spend surprisingly little time asking whether different grouping structures produce different outcomes.

That question became the focus of my doctoral dissertation. Over two years, I analyzed national mathematics achievement data to better understand how different grouping practices were associated with student achievement among Latina multilingual learners in Grade 8 mathematics. Some of what I found confirmed what I already believed; some of it completely changed my mind. Today, I want to share those findings because if you're anything like me, you're constantly wondering:

"What's actually worth changing in my classroom?"

Collaboration and grouping are not the same thing.

A Quick Note About the Research

As a note, my dissertation focused specifically on Latina multilingual learners in middle school math classrooms. While I suspect many of these findings can be applied to broader student demographic groups, my study was not designed to make those claims. Think of this article as an invitation to reflect on your own instructional practices rather than a prescription for every classroom.

The Five Student Grouping Strategies

Student grouping can be divided into five categories.

Type	Example	Pros	Cons
Whole Class	Direct instruction	All students get the same content at once	Not all students get the content Can be boring
Homogeneous Ability	Reteach group On pace group Extension group	Each group gets instruction at their level	Limited peer teaching opportunities Time intensive planning for teacher
Heterogeneous Ability	Groups intentionally containing high, medium, and low achievement students	Opportunities for peer teaching	Planning groups requires time and focus
Random	Groups created by chance selection such as cards	Opportunities for peer teaching Easy to do on the fly	Requires relinquishing some control
Student-Selected	Students choose their group mates	Students like their group mates	Stressful for shy students Easy for students to get off-task

What I Thought I Would Find

At the start of my research, I predicted that students frequently placed into heterogeneous ability groups would have the highest achievement. It seemed logical. Students would explain their thinking to each other and learn from one another. I also expected classes relying predominately on whole-class instruction to have the lowest achievement. The data disagreed.

How I Conducted the Study

I analyzed the publicly available 2024 NAEP scores for grade 8 mathematics, as well as student and teacher survey responses. I looked at results for Latinx (coded as Hispanic) students, female students, and multilingual students (coded as LEP), then interpreted the results for the intersection of those three demographic groups. My research focused on grouping structures used "often" or "always/almost always" base on teacher self-reporting.

The Finding That Changed My Mind

Contrary to what I expected, random grouping consistently showed the highest math achievement. I wasn't expecting that, and if I'm being transparent, I wasn't even sure I believed random grouping mattered that much.

My Objections

When I first started hearing about random grouping, I had concerns.

What if my two biggest distractors end up together?

What if the students that can't stand each other refuse to work together?

What if no one talks?

What if they do nothing but socialize?

Is this just another "it" trend in education that we'll abandon in five years?

These felt like legitimate concerns, but the research was compelling enough for me to change my practice.

What Happened When I Tried It

I certainly didn't have perfect groups. Some combinations still struggled. Some students still complained, while others remained frustratingly silent. However, something else happened, too. I heard more students explaining the math to each other. I answered fewer procedural questions because students were asking each other first. I was able to listen more to the mathematical thinking instead of stressing over making "perfect" groups. (Much like seating charts, there are no perfect groups.)

There's no such thing as the perfect group.

What I Still Think About

I expected whole-class instruction to have the lowest achievement. Homogeneous ability grouping was associated with the lowest achievement. This surprised me and my dissertation committee. We weren't anticipating this result. Why? Grouping multiilingual learners together often feels supportive. Students sharing a language can communicate better. With groups consisting of multilingual learners, a teacher can provide targeted instruction, as students often have comparable achievement levels. These are all understandable reasons, however, what might students lose when they are confined to working with peers at a similar level? What messages do they internalize about their math identity? How often do they hear or share problem-solving strategies and perspectives?

Three Changes I Made in My Own Classroom

⏰ I stopped spending 20 minutes making groups. Random grouping is faster.

👂 I intentionally listened for peer teaching.

💭 I became intentional about ability grouping. I still use it, but only when there are very specific instructional reasons.

So, Should We Stop Ability Grouping?

No, of course not. Like every instructional strategy, grouping structures have a purpose. Targeted intervention still matters. Small-group reteaching still matters. Enrichment still matters. The takeaway is to be intentional with grouping as it relates to the learning.

Ready to Try Random Grouping?

One reason teachers tell me they avoid random grouping is simple: it feels like more work. I've done the work for you and am sharing the exact cards I use in my own classroom (up to 33 students).

Each card includes an animal, color, number, and letter. That means one deck can instantly create groups, review teams, or station rotations without opening another app or spending precious planning time arrange groups.