Why 'nearly right' is the most expensive habit in GCSE English

There is a particular kind of student who worries me more than a struggling one.

By every usual measure, this student is doing well. Their books are full, they hand in everything, and their answers are, in the teacher’s word, fine. Yet they sit at a 6 or a 7, and they cannot work out why they are not at an 8 or a 9.

Suppose this student were your child. You would look for mistakes first, and you would not find many, because almost nothing they write is wrong. So where are the missing marks going? Almost always, the answer is that everything they write is nearly right, and “nearly right” turns out to be the most expensive habit in the subject.

A rare grade and an uncertain mark

You may know these already from our earlier article, so I will keep them brief.

A grade 9 in English Language went to 2.9% of 16-year-olds in England in summer 2025, which is fewer than three students in a hundred.^[1] Calling it “a very good grade” undersells it, because it sits in the top slice of the top fifth.

English is also genuinely hard to mark. When Ofqual measured how often examiners give the same mark a senior examiner would give, maths scored 0.96, while an English language and literature qualification scored 0.52.^[2]

Hold both facts in mind, because they meet in a surprising place.

Borderline answers are marked least reliably

Buried in that marking study is a sentence that matters more than the headline number. Ofqual found that the probability of getting the definitive mark “dips in the mark region near the grade boundaries and is highest at the extremes of the mark distribution.”^[2]

What does that mean for your child? An answer that sits on a borderline is the answer most likely to be marked inconsistently, while an answer that is clearly strong is the answer most likely to be marked correctly.

No examiner is being careless here, and the reason becomes obvious once you picture the marking. If a piece of writing is clearly excellent, any competent examiner recognises it. If it hovers between two levels, the examiner has to make a fine judgement, under time pressure, on a hot afternoon, deep into a pile of scripts that has to be finished by a deadline.

So “nearly right” costs your child twice.

First, it caps the grade, because nearly-right answers earn middle-band marks by definition.

Second, it makes the grade they do get less certain. A borderline script is precisely the script the marking system handles least reliably.

Being clearly strong earns more marks, and it also removes the doubt.

Nearly right earns marks, but never the top band

Nearly right is rarely dramatic, which is exactly why it survives so long.

– The quotation supports the concept, roughly, though it falls short of the best quotation for it. – The technique is named correctly, as in “this is a metaphor”, but the explanation of its effect could be attached to almost any metaphor in any text. – What it says about the writer’s intention is true, and also true of half the novels ever written. – The sentence is grammatically fine, and three words longer than it needs to be. – The conclusion restates the introduction rather than arriving anywhere.

Every one of those earns something, and not one of them earns the top band.

Why does the habit last so long? Nothing in the student’s experience flags it as a problem. They are not making errors, so they get reasonable marks. Teachers write “good” in the margin because the work is good. There is never a moment of failure to learn from, so the student ends up stuck on a plateau built entirely out of acceptable work.

Hard practice, honest marking

Doug Lemov, who has spent years studying what effective teachers do differently, has a technique he calls “Right is Right”. The instruction is blunt: “Do not accept partially or almost right answers; hold out for all the way.”^[3]

Half-right answers do no harm in themselves. Accepting them does, because it teaches a student that half-right is the finish line.

Holding out for the right answer is easy to turn into something harmful, so I want to be accurate about the balance.

Lemov does not argue that a student should get everything right all the time. In Practice Perfect he says the opposite: “the ideal success rate still isn’t 100 percent — if that’s the case, then the activity isn’t hard enough.”^[4]

Those two ideas fit together, and together they make the whole method:

Practice should be hard enough that your child fails sometimes. Marking should never round up.

A student who never gets anything wrong in practice is practising things they can already do. A student whose almost-right answers are quietly accepted never finds out where the standard actually is. Most of the students we meet get neither: their work is too easy and it is marked generously.

Maths shows the gap, and English hides it

In maths, “nearly right” is usually just wrong, because the numbers do not balance and the feedback is instant and hard to argue with.

English gives no such signal. A vague paragraph looks like a paragraph, and it fills the same space on the page as a precise one. From the outside, and even from the student’s own side of the desk, the two are almost impossible to tell apart.

The exam boards see the consequence. In June 2023, Question 2 of AQA’s GCSE English Language Paper 2 was an 8-mark summary question. On that question, AQA reported that “just 3% of students were awarded marks in the highest level.”^[5] Those students had not left the question blank. They had written something reasonable, and reasonable was not what the top band asked for.

Three things move a student off the plateau

You might expect the answer to be harder work, or another essay. Neither is on the list, because the list holds three other things.

1. Marking that refuses to round up. If a paragraph is nearly there, the student should be told so, and told exactly what is missing. We mark more strictly than the exam board does, on purpose, because a student who is comfortably strong against a hard standard is safe against the real one.

2. Feedback attached to the sentence rather than the essay. Instead of “develop your analysis”, the note in the margin says: this explanation would fit any metaphor in the book; tell me what THIS one does that another would not. That is a correction a student can act on immediately.

3. Doing it again. A correction that is read and not applied changes nothing, so the paragraph gets rewritten until it is right, and then the next weakness gets found.

This way of working is slower than writing another essay. It is also the only thing we have seen work reliably. We built Sophicly to do exactly that: every piece marked against the examiner’s criteria, element by element, with the correction attached and the rewrite expected.

One question to ask tonight

If you want a single, quick test of whether your child has this habit, choose a piece of work they were pleased with. Then ask them one question about it:

“Is this the best answer — or just the quickest one you could think of?”

Students almost always know the answer. The habit survives because nobody asks them the question.

If you would like to see where your child stands against the exam criteria right now, our free assessment takes five minutes and gives you a predicted grade plus the specific skills holding them back.

Sources