One Classroom, Different Levels

Walk into almost any English classroom and the idea of a single “class level” quickly begins to fall apart.

One student understands most of what the teacher says but struggles to speak. Another speaks confidently but writes with limited accuracy. Some students read quickly; others need considerably more processing time. Vocabulary, confidence, background knowledge, previous exposure to English, and familiarity with the topic can vary just as much as formal proficiency.

The obvious response is differentiation.

The difficult question is what that actually means when one teacher may have twenty or thirty students, limited preparation time, and one curriculum to teach.

Differentiation cannot realistically mean designing a separate lesson for every learner. Nor should students with lower English proficiency automatically receive simpler thinking.

Students may need different levels of access and support without needing different intellectual ambitions. That distinction can make differentiation both more manageable for teachers and more meaningful for students.

Recent evidence gives us reason to take differentiation seriously without treating it as a guaranteed solution. A 2026 systematic review and meta-analysis synthesized 45 controlled studies contributing 56 effect sizes across 18 countries. It found positive effects on average, but outcomes varied enormously across studies. The authors therefore emphasize that the effects of differentiation are conditional rather than uniform (Sedanza & Abatayo, 2026).

The variation matters. The useful question is not simply whether differentiation works. A better question is: What should we differentiate, for whom, and why?

Start With the Learning, Not the Level

When teachers see a mixed-proficiency class, it is tempting to begin by dividing students into categories.

Who is A1? Who is A2? Who is B1?

Those descriptions can be useful, but they should not become the starting point for every lesson. A CEFR level is not a complete profile of a learner: the framework describes language activity across reception, production, interaction, and mediation, alongside plurilingual and pluricultural competence (Council of Europe, 2020).

Begin instead with the learning.

Then ask a second question: What might prevent different students from accessing that learning?

For one student, the barrier might be unfamiliar vocabulary. Another may understand the idea but lack the language needed to express it. Someone else may need a model before beginning independently. Another student may need more processing time. A student who has already grasped the central idea may need greater complexity rather than more repetition.

If vocabulary is the barrier, support the vocabulary. If language production is the barrier, support expression. If task clarity is the barrier, provide a model. If the student already understands the learning, increase complexity. Do not redesign everything when changing one thing might be enough.

Different English Does Not Mean Less Thinking

Students may understand more about an idea than they can currently express in English. At the same time, language proficiency can affect how easily students access a task, participate in discussion, and demonstrate what they know. That is precisely why appropriate language support matters.

If we continually give lower-proficiency students less demanding questions, simpler ideas, and easier intellectual tasks, language support can quietly become reduced expectation.

Consider a lesson built around the question: Should wild animals ever be kept in captivity?

A student with emerging English might receive images, five essential words, short evidence statements, and a language frame such as “I think ___ because ___.” Another student might compare evidence using “Although ___, I believe ___ because ___.” A more proficient learner might evaluate competing arguments, consider exceptions, and respond to a counterargument.

Their language demands are different. The sophistication of what they can currently express in English may be different. But everyone can still participate in the same worthwhile intellectual conversation.

That is very different from giving one group the real lesson while another group spends the period completing disconnected vocabulary exercises.

Diagnose Before You Differentiate

Teachers often see the result of a difficulty before they know its cause. A student cannot answer the question. A student reads slowly. A student produces two sentences while classmates produce a paragraph. A student does not participate in discussion.

It is easy to interpret all of these as evidence that the student simply needs easier work. But the visible difficulty may have several possible causes.

A Differentiation Diagnosis

Possible barrier	What we might notice	A first response to consider
Vocabulary	Understanding improves once several key words are explained	Pre-teach essential vocabulary
Language production	Student has an idea but struggles to express it	Sentence starters or a language bank
Background knowledge	Language is manageable but context is unfamiliar	Brief context, image, or example
Processing time	Student reaches understanding but more slowly	Think time or chunking
Task clarity	Student succeeds after seeing an example	Model or worked example
Speaking confidence	Student can answer privately but avoids public response	Partner rehearsal or a lower-risk response
Reading access	Text density blocks understanding	Annotation, chunking, or visual support
Insufficient challenge	Student completes accurately with little thought	Greater complexity, transfer, or evaluation

This table is not a diagnostic test. Its purpose is to slow down a common leap: student struggles → give easier work.

Sometimes differentiation begins with better diagnosis rather than more resources.

The ACCESS Model

The ideas in this article can be brought together into a simple planning process: ACCESS. ACCESS is a practical synthesis developed here to help teachers apply the principles discussed in this article. It is not presented as a validated instructional model.

A — Aim What is the important learning that should remain common? Protect this first. If students are learning to evaluate evidence, developing English proficiency should not automatically turn the objective into memorizing vocabulary.

C — Check What is actually creating the barrier? Language? Vocabulary? Background knowledge? Task clarity? Confidence? Processing time? Or is there no barrier at all and the student actually needs greater challenge?

C — Choose What is the smallest useful adjustment? Perhaps three essential words are enough. Perhaps students need an image, a worked example, additional time, or a chance to rehearse. Sometimes substantial additional support really is necessary. The principle is not minimal support at all costs; it is no more support than is useful and no less than is necessary.

E — Extend Who needs greater complexity, independence, or transfer? Differentiation is not only about students who are struggling. Students who already understand something need somewhere meaningful to go.

S — Step Back Which support can now be reduced or removed? Scaffolds should not automatically become permanent features of learning.

S — See What happened because I made the adjustment? Did participation improve? Could the student demonstrate the learning? Was the task still intellectually demanding? Can I reduce the support next time? Did the extension actually require deeper thinking?

The final step matters because differentiation should not be judged merely by whether the teacher differentiated. It should be judged partly by what happened because the teacher did.

Support Access Before Reducing Challenge

The U.S. What Works Clearinghouse practice guide for elementary and middle-school English learners gives strong evidence ratings to intensive instruction in selected academic vocabulary and to integrating oral and written English-language instruction into content-area teaching. It gives moderate evidence to targeted small-group intervention for students struggling with literacy and English-language development (Baker et al., 2014).

The important idea is not that some students simply need easier work. They may need better access to the work.

• pre-teaching a small number of essential words;
• using an image, diagram, example, or demonstration;
• providing a worked example;
• allowing rehearsal before public speaking;
• breaking a complex instruction into manageable stages;
• providing a sentence starter or useful language structure;
• giving additional processing time;
• using a graphic organizer;
• making strategic use of students’ stronger languages where appropriate; or
• modeling what a successful response could look like.

These are possible supports, not a claim that every strategy is equally effective in every context. A 2026 systematic review of 25 empirical studies in secondary EFL classrooms found strategic use of learners’ languages across a range of pedagogical functions, while also identifying structural, ideological, and teacher-preparation constraints on implementation (Zakharova & Lasagabaster, 2026).

Change What Actually Needs Changing

A practical way to keep differentiation manageable is to identify exactly where the barrier lies. Sometimes students can work with the same material but need additional vocabulary, annotations, or visual support. Sometimes the task itself is accessible, but students benefit from first working through an example. Sometimes they understand the concept but need language support to demonstrate that understanding. Sometimes they need additional time. Sometimes they need a different route into the learning. And sometimes they are ready for greater complexity rather than more practice.

The mistake is assuming that everything must change simultaneously.

If vocabulary support solves the barrier, there may be no reason to create a different text, task, objective, and assessment as well. If students understand the reading but struggle to formulate an extended answer, rewriting the reading may solve the wrong problem.

Vocabulary Deserves Particular Attention

Vocabulary can become a hidden gatekeeper in mixed-level classrooms. Students may understand the underlying concept but be unable to access a text, question, or discussion because too many important words are unfamiliar.

The answer is not to pre-teach every difficult word. That can consume the lesson and remove opportunities for students to infer meaning from context.

Instead, identify the words students genuinely need in order to enter the learning. The What Works Clearinghouse gives a strong evidence rating to intensive teaching of selected academic vocabulary for elementary and middle-school English learners (Baker et al., 2014).

Teach those carefully. Use them repeatedly. Let students hear them, say them, read them, and use them for a genuine communicative purpose. Vocabulary becomes more valuable when it enables students to participate in the thinking that follows.

Give Students Language They Can Use

Students sometimes have an idea but cannot yet construct the English needed to express it. A small amount of language support can change the conditions for participation.

Instead of only asking “Why do you disagree?”, we might make structures available: “I disagree because…”; “The strongest evidence is…”; “One difference between ___ and ___ is…”; “At first I thought ___, but now I think ___ because…”

These are not scripts students should depend on forever. They are temporary tools.

One Possible Support Ladder

This is an illustrative progression, not a validated five-stage sequence. The research base supports the broader principles of contingent support, fading, and transfer of responsibility rather than this exact order (Ertugruloglu et al., 2023; van de Pol et al., 2010).

As students gain control of the language, prompts can become lighter. A full frame can become several phrases; several phrases can become a few keywords; eventually the student may need no prompt.

Every Scaffold Should Have an Exit Strategy

There is a danger in successful scaffolding: because the support works, we keep using it.

A sentence frame helps, so the student receives another. A simplified text helps, so texts continue to be simplified. Translation helps, so translation is always provided. Teacher prompting helps, so the teacher continues prompting.

Each decision may make sense in isolation. Together they can gradually reduce how much the learner needs to do independently.

We might think of this as differentiation debt: what can happen when support is continually added but rarely reconsidered or removed. The term is used here as a practical description, not as an established research construct.

The underlying concern, however, is well established in scaffolding literature. A major review identified contingency, fading, and transfer of responsibility as central characteristics of scaffolding, while a later review of bilingual secondary education also identified fading and increased responsibility or autonomy among recurring characteristics (Ertugruloglu et al., 2023; van de Pol et al., 2010).

A model becomes a checklist. A sentence frame becomes keywords. Teacher-led practice becomes partner practice. Partner practice becomes independent work. A heavily annotated text becomes a lightly annotated text. Successful differentiation can sometimes be measured by something no longer being necessary.

Use Grouping With a Purpose

Mixed-proficiency classes naturally raise questions about grouping. Should similar-level students work together? Should different proficiency levels be mixed?

There is no single answer because grouping should depend on what students are being asked to do.

Sometimes a teacher needs a short targeted group addressing a particular literacy or language need. The What Works Clearinghouse gives moderate evidence to targeted small-group intervention for elementary and middle-school English learners struggling with literacy and English-language development (Baker et al., 2014). That evidence should not be stretched into a claim that every mixed- or same-proficiency grouping arrangement improves learning.

At other times, mixed-proficiency groups may be useful because a task benefits from different strengths and perspectives. The important point is that grouping should have an instructional reason.

Students should not permanently become the “low group” and “high group.” A student’s performance and support needs can change across language activities, tasks, and time.

Not Everything Needs to Be Differentiated

This may be one of the most useful things teachers can hear.

You do not need three worksheets for every lesson. You do not need three versions of every reading. You do not need three homework assignments simply because three proficiency levels exist in the room.

Sometimes the same clear teacher explanation benefits everyone. Sometimes one image makes an idea accessible. Sometimes everyone can attempt the same question and differentiation occurs through prompts and feedback. Sometimes support can simply be available to anyone who wants it. And sometimes the original task is already accessible.

Differentiation should respond to a genuine learning need. It should not become a ritual.

The 2026 meta-analysis reinforces the need for caution: although effects were positive on average, results varied widely across contexts and studies. “Differentiated” does not automatically mean “better” (Sedanza & Abatayo, 2026).

The 60-Second Differentiation Check

That may be enough. Differentiation does not always require another worksheet, another lesson plan, or another hour of preparation.

Build a Support Menu

Another way to make differentiation sustainable is to stop making every support teacher-assigned.

• key vocabulary;
• visual references;
• sentence starters;
• models of successful work;
• bilingual keywords where appropriate;
• checklists;
• optional challenge questions;
• graphic organizers;
• opportunities to rehearse before presenting.

Students can increasingly choose what they need. The long-term goal is not for the teacher to personally customize every moment of learning. It is for students to become increasingly capable of recognizing what they are trying to do, what is getting in the way, and which available support could help them continue.

Differentiate the Question Before the Text

Teachers sometimes spend considerable time producing multiple versions of the same reading. Sometimes that is necessary. But often everyone can engage with the same central material if access support is available.

What changes can be the question. After reading the same text, students might consider: What happened? Why did it happen? Which explanation is most convincing, and what evidence supports your judgment?

These are examples of different response demands, not permanent “easy, medium, hard” tracks.

Give Thinking Time Before Speaking

Some students may understand a question but need more time to formulate an answer in English. Changing “teacher question → immediate answer” to “teacher question → thinking time → partner rehearsal → whole-class response” can give them a better opportunity to prepare the language required to participate.

The intellectual task has not necessarily been simplified. The conditions for participation have changed. This is a practical strategy rather than a universal claim that rehearsal will improve achievement in every context.

Language and Content Can Develop Together

English learners do not necessarily need to master English first and encounter interesting academic content later.

The What Works Clearinghouse gives strong evidence to integrating oral and written English-language instruction into content-area teaching for elementary and middle-school English learners (Baker et al., 2014). Research on bilingual secondary education similarly treats language scaffolding as support for engaging with disciplinary content through a language learners are still developing (Ertugruloglu et al., 2023).

A science investigation can develop comparison, prediction, evidence, and causal language. A history lesson can develop the language of interpretation and argument. A story can generate prediction, explanation, and evaluation. A mathematics problem can require students to explain reasoning, justify choices, and compare solutions.

Differentiation Should Not Lower the Ceiling

Differentiation is often discussed primarily in relation to students who need additional support. But mixed-level teaching also requires us to consider students who are ready to go further.

Possible ways to increase intellectual demand include asking students to describe, explain, compare, evaluate, transfer an idea to a new context, or create. These are not presented as a fixed hierarchy: the difficulty of any task depends on its content, the learner’s knowledge, and what the task actually requires.

Same Learning, Different Access

Consider a class investigating: Should zoos keep wild animals?

The common learning goal might be: Use evidence to develop and justify a position.

• Access support: five essential words, images, and short evidence statements.
• Language support: an optional sentence frame such as “I believe ___ because ___.”
• Standard task: use evidence to explain and justify a position.
• Greater challenge: identify the strongest counterargument, evaluate its evidence, and respond.

Same intellectual conversation. Different amounts of access, structure, independence, and challenge.

Keep Differentiation Sustainable

Teacher time and attention are finite. A model of differentiation that requires teachers to continually create multiple texts, tasks, assessments, resources, and homework assignments may become unsustainable.

Before spending forty-five minutes creating three versions of a worksheet, ask whether a few carefully chosen supports and one meaningful extension could address the actual barriers. Sometimes more preparation will be necessary. But complexity should earn its place.

What About AI?

Generative AI changes this conversation in an interesting way. It can produce multiple versions of a text or worksheet almost instantly. A teacher can ask an AI system to rewrite a task at A1, A2, and B1, and seconds later three versions may appear.

That certainly makes one form of differentiation faster. But it does not tell the teacher whether creating three versions was the right educational decision.

If the barrier is five unfamiliar words, rewriting the entire text may unnecessarily reduce complexity. If the student understands the concept but cannot express an argument, simplifying the content solves the wrong problem. If the student needs more challenge, generating additional recall questions simply creates more work.

AI may therefore be more useful after the teacher has diagnosed the need. A teacher might ask it to identify potentially blocking vocabulary, draft optional sentence starters, suggest an alternative explanation, generate an extension question requiring evaluation, or propose a way to fade a scaffold used in the previous lesson.

The teacher still makes the educational decision. AI helps produce or adapt the resource.

Do I Actually Need to Differentiate This?

Sometimes the best differentiation decision is not to differentiate.

A Differentiation Planning Card

The final question may be the most important. Differentiation should not be judged merely by whether the teacher differentiated. It should be judged partly by what happened because the teacher did.

One Classroom Does Not Require One Path

Mixed-level teaching will never become perfectly tidy.

Students develop at different speeds. Their strengths do not fit neatly into proficiency labels. Their abilities can vary across reading, writing, listening, speaking, interaction, and mediation. Background knowledge matters. Confidence matters. The nature of the task matters. The support that works today may be unnecessary next month.

That is normal.

Recent evidence gives us reason to take differentiation seriously without pretending that every differentiated strategy works equally well everywhere. The 2026 synthesis found positive effects on average but very substantial variation across studies and contexts (Sedanza & Abatayo, 2026).

Differentiation is not a program. It is not three worksheets. It is not permanently dividing students into high, middle, and low groups. And it is not asking one teacher to design thirty separate lessons.

It is a process of making better instructional decisions in response to learner differences.

That leaves teachers with a manageable set of questions: What is the important learning here? What is preventing this student from accessing it? Which part of the learning actually needs adjustment? Am I supporting the student’s language—or reducing the student’s opportunity to think? What support is this student ready to lose? What changed because I differentiated?

Differentiation becomes far more practical when we stop trying to create a different lesson for every level.

A mixed-level class can work toward meaningful shared learning while students receive different amounts of language support, structure, time, independence, and challenge.

References