Teaching A1 and B2 in the Same Room: A Working Guide to Mixed-Level Classes

A placement test puts eighteen students in a B1 class. By the end of week one you have found that four of them cannot reliably produce the past simple and two of them read news articles in English for fun. Nothing has gone wrong. That is what a B1 class is.

Mixed levels are not an edge case you occasionally survive. They are the normal condition of group teaching, and the difference between a teacher who copes and a teacher who does not is a handful of design decisions made before the lesson starts. This post lays those decisions out.

The material here comes from the classroom management and lesson planning modules we teach in our own courses, so it reflects one coherent approach rather than a survey of the field. Where a technique has a real trade-off, we say so.

Every Group Class Is Mixed Level

Unless you work somewhere with strict placement testing and the willingness to move students mid-course, you will teach mixed levels. It is most pronounced in three settings: community programs, in-company courses, and small language schools, all of which enroll on availability rather than on ability.

Even a well-tested group spreads. A class labelled B1 will hold strong A2 learners and weak B2 learners on day one, and the spread widens over a term because students attend unevenly, study unevenly, and start from different first languages. A Spanish speaker and a Japanese speaker at nominally the same level are not carrying the same difficulties, and no single label captures that.

The practical consequence: plan for a range, not a point. If your lesson only works when everybody is at the level on the enrollment form, it will not work.

The Default Failure: Teaching to the Middle

The instinctive response to a spread is to aim at the average. It is also the one response that fails everybody at once.

Aim at the middle and your strongest students finish early, disengage, and start treating the class as a scheduling obligation. Your weakest students spend the hour a half step behind, understand roughly a third of what happens, and learn that asking for help costs them face in front of the group. The middle third gets a lesson pitched correctly and is the only part of the room you are actually teaching.

The fix is not to slow the class down. Slowing down keeps the same failure and adds boredom. The fix is to change what students produce rather than what they receive.

Same Input, Different Output

This is the whole idea, and it is worth stating plainly: differentiate the task, not the class.

Everybody reads the same article. A1 answers five comprehension questions. B2 summarizes it and gives an opinion. Everybody looks at the same picture. A1 writes three sentences. B2 writes a paragraph comparing what the picture shows with their own country.

Two things happen. First, the class stays a class. Shared input means shared reference points, and students can talk to each other about the same thing afterwards. Second, you prepare one thing instead of three.

The same move works on listening, which is where teachers most often assume they are stuck. You do not need three recordings at three speeds. Play the audio once for everyone, then ask beginners what the speaker talked about and ask advanced students what the speaker's opinion was and whether they agree. One recording, two levels of demand, no extra preparation.

The Cheapest Differentiation in the Room Is the Word "Or"

Most teachers hear "differentiation" and picture building three versions of every worksheet. That is not sustainable across a full timetable and it is not necessary.

The lowest-cost differentiation is spoken. It costs one sentence of instruction and no preparation at all:

- "Answer in a word or a sentence."

- "Write a list or a paragraph."

- "Everyone does questions one to five. If you finish, do six to ten."

- "Tell me about your hometown. Five sentences, or a paragraph with two comparisons, or two paragraphs on how it has changed."

Ranked from least to most preparation time, your options are roughly: spoken output instructions, then extension questions added to material you already have, then a supplementary text for the advanced students, then a brand new worksheet for the lower level. The first two cover most of what you need. The last one is where teachers burn their evenings for very little extra return.

Building a Tiered Task in Four Steps

When you do want a structured tiered activity rather than an improvised one, the sequence is short:

1. Identify the core topic everyone will study. One topic keeps the class cohesive.

2. Create two or three versions of the task at different levels of demand.

3. Either assign versions by level or let students choose their own.

4. Bring the class back together to share results.

Step four is the one people skip and it is the one that makes the activity feel like a single lesson rather than three parallel ones. If the room splits into tiers and never reconverges, weaker students learn that they are in the slow group, which costs you more in confidence than the task gains in accuracy.

On step three, letting students choose is often better than assigning. Self-selection is fast, avoids public labelling, and is usually accurate. Students know where they are. The occasional student who picks the easy version to coast is a smaller problem than a room where everybody has been publicly graded.

Pairing: Strong With Weak, but Not Always

Pairing a stronger student with a weaker one is the standard advice and it does work, for a specific reason: the stronger student has to explain, and explaining is where passive knowledge turns into active knowledge. The weaker student gets attention you cannot personally provide in a class of twenty.

Here is a version of that pairing done properly. The teacher gives each student a role rather than a shared instruction: the A1 student describes the picture out loud, and the B2 student writes down what the A1 student says and helps when they get stuck. The A1 student practices speaking at their level. The B2 student practices listening, writing, and supportive questioning, which are real B2 skills and not charity work. Both are producing language. Neither is waiting.

That last point is the test to apply. A mixed pairing works when both students have a task that develops them. It fails when the stronger student becomes an unpaid teaching assistant, which is what happens when the instruction is simply "help your partner".

And it should not be your only pairing. Strong students sometimes need to be pushed by other strong students. Weaker students sometimes need the confidence of working with someone at their level, where they are not the one who is always behind. Vary the pairing based on the goal of the activity, and rotate partners between lessons rather than letting students settle permanently next to the same person.

Group Size: Three or Four, Never Five

Groups of three or four are the working range. Above four, somebody always hides, and the student who hides is usually the one who most needed to speak.

Use pairs when you want maximum speaking time per student. Use groups of three or four when the task genuinely needs collaboration. Use individual work when students need thinking time before they say anything out loud, which is the most commonly skipped stage in mixed-level classrooms.

If a group of four contains a student who has not spoken, the fix is structural rather than personal. Assign roles: timekeeper, note-taker, reporter, question-asker. When everyone has a job, nobody can hide, and you have not had to single anybody out in front of the group.

Think-pair-share exists for exactly this problem. Thirty seconds of individual thinking, then a partner, then the class. The individual stage is what lets a slower processor arrive at the pair stage with something to say.

The Fast Finisher

The B2 student who finishes every activity in half the time is not a discipline problem. They are an unplanned-for problem, and there are three common responses, two of which are bad.

Telling them to wait quietly teaches them that speed is punished. Sending them to help others every single time turns them into staff. The workable answer is an extension task that pushes beyond the base activity: "now write a paragraph using all five words", "now write your own example", "now argue the opposite position". Peer support is fine occasionally. It should not be their only option, because it stops being their lesson.

Build the extension into the material once and you have it forever. Any worksheet gains a final question that no one is required to reach.

The Student Who Says Nothing

A silent student is rarely lazy. They are usually scared, confused, embarrassed, or operating under a cultural norm about speaking up that you have not asked about.

The useful question is not "why are they not trying" but "what is stopping them". The answer changes the response. A confused student needs the instruction repeated differently. A scared student needs a lower-stakes format, which usually means pairs rather than open class. An embarrassed student needs to not be called on cold in front of eighteen people.

One structural note about silence: if a conversation class is quiet for more than five minutes, that is a signal about your activity design, not about the students. Vague instructions produce silence. "Discuss the topic with your partner" reliably produces two people nodding at each other. "Tell your partner three reasons you agree or disagree, and be ready to report one of them" produces talking, because it names a quantity, a stance, and a consequence.

The Student Who Says Everything

Every class has one. It is tempting to treat the over-talker as the problem, but a student who talks constantly is an engaged student, and engagement is harder to create than to channel.

The real cause is usually structural. Open-class questions reward whoever is fastest, and the fastest student is the same person every time. Restructure so that speaking is distributed by design: pair and group work with assigned roles, nominated turns rather than raised hands, time limits on each speaker.

What not to do is tell a student they talk too much. It is a personal criticism for a structural failure, and it usually costs you the most willing participant in the room.

Monitoring and Correction When Levels Differ

Walking around the room is not monitoring. Monitoring means deciding in advance what you are listening for: a specific recurring error, creative use of the target language, students who are stuck, students who are off task. Wander without a target and you will collect nothing usable.

Correction strategy should switch with the activity, and in a mixed-level room this matters more, because the same error means different things at different levels.

- During controlled practice, where accuracy is the point, correct immediately.

- During fluency work, note errors and address them after the activity. Interrupting a weaker student mid-sentence to fix a preposition is the fastest way to make them stop volunteering.

Carry a notepad or use your phone. Collect four or five errors during the activity, then put them on the board afterwards as anonymous sentences for the class to fix. Nobody is identified, the strong students get a puzzle, and the weaker students see their own error corrected without having been the example.

Implicit correction, where you guide a student to fix it themselves, beats explicit correction almost every time, because students remember the corrections they worked out. It is also slower, which is the trade-off, and in a large class you will not have time to do it for everything.

What a 60-Minute Mixed-Level Lesson Looks Like

Concretely, with the differentiation marked:

Lead-in, 5 minutes. A picture prompt or a personal question. No level dependency. Everyone can answer at their own depth.

Input, 15 minutes. One reading text or one recording for the whole class. Gist task first, one simple question. Then a detail task with tiered questions: everyone does the first set, the second set is harder.

Language focus, 10 minutes. Pull the useful language out of the text. Meaning first, then form, then pronunciation.

Practice, 15 minutes. Pair or group task with roles assigned. Output instruction differentiated out loud: a list or a paragraph, three sentences or five.

Feedback, 10 minutes. Delayed correction from your monitoring notes, on the board, anonymous.

Cool-down, 5 minutes. Exit ticket tied to the objective, so you find out who actually got it.

Every stage above is either level-neutral or differentiated by output. None of it required a second set of materials. The total preparation is one text and one decision about what the tiers are.

Build in a buffer. Transitions eat time, instructions eat time, and handing out paper eats time. If your stages add up to exactly sixty minutes, the lesson is already over-planned.

Which Jobs Put You in a Mixed Room

Worth knowing before you sign, because the setting predicts the spread.

Adult language centers and in-company courses are the most mixed, because enrollment follows the corporate calendar and the client's staff list rather than any test. Community and government programs are next. Private academies with real placement testing and enough students to run parallel classes are the least mixed, which is one of the quiet advantages of a larger school.

As a rough indicator of the shape of the market, of the 155 listings in our current job snapshot, 130 are in-person, 22 are online, and 3 are hybrid, and 117 are full-time against 29 contract, 8 part-time, and 1 freelance. That is a single day's view of one board rather than a survey, so treat it as an order of magnitude. What it does tell you is that most advertised roles are full-time classroom work in a physical room, which is precisely the setting where you cannot quietly re-level a class the way a one-to-one tutor can.

Online one-to-one work removes the problem entirely and replaces it with a different one: you carry the full cognitive load of adapting to each student with no group energy to lean on. Our comparison of online versus in-person teaching covers that trade-off.

What to Ask a School Before You Sign

Four questions, and the answers tell you how much differentiation work you are being handed:

1. Do you run placement tests, and can students be moved between classes mid-term?

2. What is the typical class size?

3. Are materials provided, or do teachers build their own?

4. How many different levels will I teach in a week?

A school that tests, places, and moves students has absorbed some of this work for you. A school that enrolls anyone into any class and provides no materials has handed you all of it, which is fine if the pay reflects it and worth knowing if it does not. Our guide to contract red flags covers the rest of what to check before signing, and the country guides give you the local norms for school type and class size.

Bottom Line

Mixed levels are the default, not the exception, and the response that fails is aiming at the average. Keep the input shared so the class stays a class, vary what students produce, and let a spoken instruction do the work that a stack of differentiated worksheets was going to do.

The rest is structure. Roles stop students hiding. Groups of three or four stop students hiding. Extension tasks stop your fastest student from checking out. Delayed correction stops your slowest student from going quiet. None of it requires more preparation time. It requires the decisions to be made before the lesson rather than during it.

Next Steps

Go deeper on the techniques in our Classroom Management for ESL Teachers course, plan faster with the Lesson Planning Masterclass, read our companion post on planning a lesson in 20 minutes, browse open ESL jobs, or submit your resume so schools can find you.