What we keep after we forget

Prior-year content, and how much to include

Someone recently asked me: “How much of last year’s content should we put in this year’s annual assessment?” They were writing papers for a MAT and needed a number. Twenty per cent? Forty?

I told them: it depends. That sounds evasive, but it isn’t. The reason it depends is itself the most useful thing to understand about how assessment works in different subjects. Let me tell you about another school that asked a different version of the same question.

A humanities mystery

Some years ago, a Head of Curriculum at a large MAT invited me to visit. Their humanities progression model had run into trouble. Students across the trust were sitting termly assessments, and when you plotted the scores across the year you saw rises in attainment followed by falls.

What was going wrong? It wasn’t that the questions were getting harder. It was that the termly assessments had started incorporating questions from earlier terms: topics studied months ago, in some cases close to a year ago. And the students were forgetting.

Of course they were.

Forgetting is not something we like to talk about as teachers. It feels like an accusation, a sign that something has gone wrong in the classroom. But it is central to how learning works, and working out how we feel about it matters a great deal for how we design assessments.

Why maths works differently

Imagine the same exercise in maths: Year 9 students sitting a paper that includes Year 7 content. Most would successfully answer them. Ratio, place value, multiplicative reasoning, the handling of fractions: these ideas get revisited constantly across the secondary curriculum, not necessarily through revisiting in an identical form but rather because later maths depends on them. You cannot teach simultaneous equations to a student who has forgotten how to solve linear ones. The curriculum does the work of keeping prior content warm.

Maths is a hierarchically structured subject. Knowledge is revisited, in increasingly complex ways. Schools teach broadly the same content in broadly the same order. When things go well, forgetting doesn’t really happen, because the curriculum itself prevents it.

Now try the same thing in history, or English literature, or geography. Year 9s may have studied the Tudors in Year 7 and not looked at them since. They may have read a particular novel in Year 8 and never revisited it. These are cumulative subjects: you take a walk through a landscape, and the places you visit are not systematically revisited. Forgetting happens. Not because anyone has failed. Because the structure of the subject invites it.

What should remain

This is where the interesting question appears. If forgetting is structural in cumulative subjects, what do we actually want students to retain?

Christine Counsell, talking about history, uses a lovely image: the residue left in a sieve. The fine grains wash through; what remains is the essence. In history, that might be a sense of period, a feel for how causes accumulate, a grasp of the shape of historical argument, or enough to better understand the period that follows it. In literature, it might be an instinct for how form creates meaning. In geography, a feel for how physical and human landscapes interact in a particular context.

The specifics, whether dates, quotations or place names, mostly don’t survive. We should be relaxed about that, provided we have been intentional about what we wanted to remain in the first place. That last clause does a lot of work. Allowing forgetting is not the same as planning for it. A department that has not articulated what the residue should be has simply abdicated the decision.

This is what separates a good cumulative curriculum from a drifting one. Not that nothing is forgotten, but that someone has thought clearly about what shouldn’t be.

Back to the primary school leader’s question

So: how much of last year’s content belongs in this year’s test?

In a hierarchical subject like maths, a great deal of it because later learning depends on earlier learning, and if students have forgotten the prior content, that is diagnostically important. You want to know.

In a cumulative subject, the question is different. You don’t ask how much of last year’s content belongs in the paper. You ask what you wanted students to retain from last year, and you test for that. Often this means assessing the disciplinary skills that were developed last year, but in a novel setting that they have recently studied. The retention you care about travels.

If, on the other hand, you decide you do want to assess whether pupils can still tell you the causes of the English Civil War eight months after studying it, you can. This is a deliberate choice about what the residue is. And be prepared for the scores to come down, because students will forget a little.

A practical prompt

So, as you bring staff together in your school to plan curriculum and assessment, think about this question from the perspective of each subject: What do we expect to remain from last year once the specifics have faded?

If you can answer it, designing the assessments becomes tractable. If you cannot, no amount of tinkering with what percentage of prior-term or prior-year content to include will fix the problem, because the problem isn’t getting the percentage right. It is deciding what the assessment is for.

The curriculum is the progression model, people sometimes say. In a hierarchical subject, that is very nearly enough to describe our desires for learning and attainment. In a cumulative subject, you also need a theory of what remains.