For eleven years, your child has written for someone who knows them.
Their teacher knows how they think. Knows what they meant by that clumsy sentence. Knows they read the whole book, so a thin paragraph about it gets read generously. When a teacher marks a piece of work, all of that quietly comes with it.
Then, one morning in May, your child writes for someone who has none of it.
The examiner has never met them and will never meet them, and your child’s script is one of possibly hundreds they will work through in that series. It arrives inside a marking window in which AQA alone handles more than seven million scripts.[6]
Everything that teacher knew is unavailable. All that stranger has is what appears on the page, and a document setting out what earns marks.
Those two audiences want overlapping but different things. Most students never find out how different until it is too late to matter.
A question worth asking tonight
Here is something you can try, and it takes about a minute.
Ask your child how their English exam is marked. Not what is on it. How the marks are given out. What is the examiner looking for in a paragraph of analysis? What separates an answer near the top from one in the middle? How many marks are going for the writing, as opposed to the reading?
We ask a version of this of every student who joins us, and the answers are remarkably consistent. Strong students, weak students, students at excellent schools. They can tell you the topics. They can tell you what their teacher tends to say. Very few can describe, with any precision, what the examiner is actually rewarding.
This surprises parents. It stops surprising you once you notice who the audience has been all year.
Working to please a person
When a teacher is the reader, a student learns to satisfy the teacher. That is sensible behaviour, and it usually works. Teachers are consistent, they explain what they want, and they mark generously when they can see the understanding behind an awkward sentence.
So a student builds up an internal picture of “what good work looks like” that is really a picture of what this teacher responds well to. Highlighted quotations. Longer answers. The technique named. A neat conclusion.
Much of that overlaps with what the exam rewards. Not all of it does, and the parts that do not are invisible while a teacher is doing the marking.
Then the audience changes without warning.
You cannot improve work you cannot judge
The most useful idea here comes from an Australian researcher, Royce Sadler, in a 1989 paper that has shaped how assessment is understood ever since.[1]
His argument is simple to state. For a student to improve their own work, they have to be able to judge its quality while they are producing it. That requires holding a picture of what good work looks like that closely matches the one the marker holds. Without it, the student is writing hopefully. They can only find out whether it worked by handing it in and waiting.
A student in an exam hall gets no chance to hand it in and wait. They have to make that judgement themselves, in the moment, about their own paragraph, alone.
The part that surprises most parents
The obvious response is to print off the mark scheme and give it to them.
Every school in the country has already done this. The criteria are published, displayed on classroom walls and handed out before mocks. The students who cannot describe what the examiner wants have almost all been shown the document.
Sadler explains why that is not enough, and it is the most useful thing in this article. Standards are partly tacit. They cannot be fully captured in words.[1] A mark scheme is a summary of a judgement rather than the judgement itself. Phrases like “perceptive”, “judicious” and “convincing” mean something precise to an examiner who has marked a thousand scripts. To a fifteen-year-old reading them cold, they are just approving adjectives.
Being handed the criteria is like being handed the rules of a sport you have never watched. Nothing in the wording tells you what the thing actually looks like.
What does work
Researchers have tested this directly. Chris Rust, Margaret Price and Berry O’Donovan ran a study on exactly this problem, and found that publishing the criteria produced little on its own.[2]
What produced measurably better performance was different: students were given real pieces of work, asked to mark them against the criteria themselves, and then made to discuss and defend the marks they had given. Arguing about a judgement is what transfers the part of the standard that cannot be written down.
The wider evidence agrees. Teaching students to monitor and evaluate their own work is one of the strongest and cheapest things education research has found, worth around seven months of additional progress — and it works best when taught inside the subject itself rather than as a separate study-skills lesson.[3][4]
Which means this is learnable. It just cannot be learned by reading.
Why it shows up worst under pressure
There is one more layer, and it explains the results that seem to make no sense.
Psychologists have spent years studying why people underperform in high-pressure moments. The finding that matters here concerns working memory, the small amount of information a person can actively hold in mind at once. Under pressure, worry occupies part of that space. Whatever the task needed it for, it no longer has it.[5]
The uncomfortable detail is who this affects most. In these studies, it is often the people with the greatest working-memory capacity who suffer, because they rely on strategies that lean on holding a lot in mind. Take the capacity away and their method goes with it. Students using simpler, more automatic approaches are less disturbed.[5]
Apply that to an exam hall. A student whose entire method is remember what the teacher said we should do is running on the exact resource the pressure is eating. A student who understands what is being rewarded has something sturdier to fall back on, because it is understanding rather than recall.
This is where the disappointing results come from. Not from lazy students. Often from strong ones, in good schools, who did everything asked of them and had nothing underneath it when the room went quiet.
Two things must be said honestly
First, this research was not done in exam halls. Much of it involves maths problems in laboratory conditions, and the step from there to GCSE English is an inference. We think it is a reasonable one, and we would rather label it than dress it up.
Second, you will see parents on social media every August whose children did well all year and came home with a grade five or six. That pattern is real and widely observed, but we could not find a sound British statistic measuring it, so we are not going to invent one. Treat it as something to recognise, not something we can count.
What to do about it
Start with the question at the top of this article. If your child can only answer it in general terms — “you have to analyse the language”, “you need quotations” — that is worth knowing now rather than in August.
After that, the fix is not more reading of the criteria. It is using them. Marking a real answer against the real standard. Comparing a strong response with a nearly-strong one and being able to say precisely what makes the difference. Doing that often enough that the language of the mark scheme stops being someone else’s vocabulary.
That is how our marking works, and it is deliberate. Every piece a student writes comes back marked against the examiner’s criteria, with the specific things that gained and lost marks named. Over a term, a student stops guessing what a marker wants and starts knowing.
On the morning in May when your child writes for a stranger, that knowledge is the only thing that travels with them.
