Differentiation without 27 worksheets

The standard primary differentiation model goes like this. Three versions of each worksheet. Bottom group, middle group, top group. 'Must, should, could.' Sometimes more — a SEND version with extra scaffolding, an extension version with deeper questions. By the end of a planning afternoon, you've made five versions of one piece of paper.

Then you discover, in the lesson, that children placed in the wrong group can't access their worksheet. You hand out swap-overs. A child who finished the 'should' is now on the 'could' but it assumes things they don't know. Another child can't do the 'must' because she misses one prerequisite skill. The whole carefully-planned three-tier system is held together by you constantly intervening.

This model is killing teachers and isn't actually helping children much. Here's why, and what works better.

What three-tiered worksheets assume

The three-tier model assumes:

- Children's ability is broadly stable across topics. - A child who needs the easier worksheet for one topic will need it for all. - Three groupings is enough granularity to capture the variation in a class. - The challenge in 'should' and 'could' versions is in the maths/content, not in the language or layout. - Children correctly placed in groups stay engaged because they're working at the right level.

Almost none of these assumptions hold up well.

Children's ability varies enormously by topic. A child who's strong in arithmetic might be weak in geometry. A child who finds writing hard might be brilliant at oral comprehension. The 'top group' for one lesson is the 'middle group' for another. But three-tiered planning encourages teachers to assign children to FIXED groups for whole units or terms, which means individuals routinely end up at the wrong tier.

Even when groups are reset by topic, three is rarely enough granularity. A typical Y4 class spans reading ages from 6 to 11. Three buckets across that range still means each bucket contains a 2-year span of ability. A child at the top of the 'middle' is usually a long way from a child at the bottom of the same group.

And the differentiation often happens through reducing volume rather than adjusting cognitive demand. The 'lower' version has fewer questions, simpler numbers, more space for handwriting. The actual thinking required is the same shape — just smaller. That's not differentiation, that's load reduction.

What works better: differentiate by depth, not by sheet

Same task. Different responses.

A maths lesson on perimeter, instead of three sheets, might have one task: 'Find as many shapes as you can with a perimeter of 20cm. Sort them by area.'

A struggling child can find a few simple rectangles. A confident child can find dozens, sort them, notice patterns, prove things, generalise. Same starter, vastly different outputs.

A writing lesson, instead of three writing frames, might have one prompt: 'Write the moment your character realised something had changed. Use the techniques we've been practising this week.' A child working at age 7 produces a paragraph using two of the techniques. A child working at age 11 produces a page using all of them, plus some they invented.

Open-ended tasks differentiate themselves. The teacher's job becomes less about sorting children into pre-made tiers and more about giving good feedback in the moment, pushing each child slightly harder than where they came in.

This is sometimes called 'low-floor, high-ceiling' planning. The floor is low enough that everyone can start. The ceiling is high enough that no-one is bored. The differentiation happens in the doing, not in the worksheet.

What works better: differentiate by support, not by task

When everyone is doing the same task, you can move differentiation into the SUPPORT each child gets, rather than into different work.

This means deciding, in advance, who needs:

- A pre-teach (5 minutes before the lesson, on the key concept) - An adult next to them for the first 10 minutes - A vocabulary card for tier-2 words they'll meet - A visual reminder (number line, sentence starters) on the desk - A modified expectation (write 5 sentences instead of 10, show your method without solving the whole problem) - Frequent check-ins from the teacher - Nothing — they're fine.

Notice that all these children are doing the SAME TASK. The differentiation is in the structure around the task, not the task itself. This avoids the social isolation of being on a 'different sheet' and means the high-ability child can see what an excellent answer looks like (because the strong children in the room are working on the same thing).

This is closer to the 'mastery' approach popular in maths but it works across subjects. Same content, varied support, ambitious for everyone.

What about SEND children?

Children with significant SEND often need genuinely different work — that's accurate. But this is a different conversation from typical-spread differentiation. SEND provision is about meeting specific identified needs (working memory, processing speed, reading age significantly below year group). It usually involves an EHCP or SEND support plan that defines what's needed.

The mistake is to treat the bottom 20% of a class — children who are 'just' lower attainers, not SEND-identified — as if they need fundamentally different work. They mostly don't. They need the same task with more scaffolding and more frequent feedback. They need to see the high-quality work the rest of the class is producing, not be quietly removed from it onto a thinner sheet.

This is hard for teachers to accept because it feels less safe. Putting a struggling child on an easier worksheet is visible — you can show your evidence of differentiation. Inviting them into the main task with extra support is invisible — there's no easy artefact to show in a book scrutiny. The visible version feels more accountable. The invisible version actually works better.

What about extension?

The 'top' children are the most quietly underserved by tier-based differentiation, in my experience.

The 'extension' worksheet is usually 'the same task but harder numbers' or 'one tricky question at the end.' Neither is real extension. A genuinely able child needs deeper, more open-ended challenge — not just more of the same.

Better extensions:

- 'Find a counterexample.' - 'Prove this always works.' - 'Show two different methods and explain when each is better.' - 'Write a question that would catch out someone in this class.' - 'Where else might this idea apply?'

These work because they push thinking, not pace. A bored top-end child doesn't need to be sent to next year's work. They need the same content treated more rigorously.

The takeaway

Stop making three sheets. Make one good task that has a low floor and a high ceiling. Plan the SUPPORT each child needs around that task — pre-teaches, adult deployment, scaffolds, modified expectations. Save the planning hours and put them into thinking about what good looks like and how to give individual feedback.

This is harder in some ways than three-tier planning. You can't show your differentiation as easily in a planning folder. You have to be more attuned to children moment-by-moment in the lesson.

But it's massively easier in others. You're planning ONE task per lesson, not three. You're not making swapover sheets when you've grouped someone wrong. You're not trapped in pre-made groups that don't match this week's strengths.

And — most importantly — it works better. Children stay in the same task, can see each other's work, feel part of the same lesson. That sense of 'we're all doing this together' is itself a learning condition. Three-tier planning quietly removes it.