Metacognition for Kids: Four Questions Worth Eight Months of School
Metacognition for kids means teaching a child to think about their own thinking: how sure they are, what they tried, where exactly they got stuck. The Education Endowment Foundation rates it at about eight months of additional progress a year, across 355 studies. Four questions, asked in ninety seconds after homework, are enough to start.
The Education Endowment Foundation reviews every credible study it can find on what actually helps children learn, then converts the results into a single number: how many extra months of progress this gives an average pupil in a year.
Many approaches on that list add a few months at most.
Metacognition and self-regulation sits at +8. Its review was last updated in May 2025 - 355 studies in all - and the figure went up, from seven months to eight. That is unusual: education findings more often shrink as more evidence arrives.
It’s also one of the cheapest things on the list. It costs nothing but the questions you ask.
What it actually means
“Metacognition” is a terrible word for a simple thing: thinking about your own thinking. Knowing what you know. Noticing when you’re lost. Being able to say what you tried and why it didn’t work.
Here’s the gap it fills. A child doing maths homework is running one process - solve the question. A child who’s learned metacognition is running two: solve the question, and watch themselves solving it.
The second process is what lets them catch an answer that can’t be right. It’s what makes them stop after two minutes of going nowhere and try something else, instead of grinding the same wrong method for twenty. It is the difference between a child who gets stuck and a child who knows they’re stuck.
Almost no homework asks for it. Worksheets ask what the answer is. They never ask how sure you are, or what you tried first.
The four questions
These are the four we use. They take about ninety seconds and work on any subject.
1. “How sure are you, from 1 to 5?”
Ask before you check the answer. You are looking for the mismatch: a confident 5 that’s wrong, or a nervous 2 that’s right. Both are more useful than the mark.
A child who’s right-and-sure has learned it. A child who’s right-and-unsure got lucky, or has it but doesn’t trust it - different problems, different fixes. A child who’s wrong-and-sure is the one to pay attention to, because they have no internal signal telling them to check, and that is exactly the child who loses marks they “knew”.
2. “What did you try first?”
This surfaces the approach, not the answer. Often the first attempt was reasonable and abandoned too early. Sometimes there wasn’t one - they waited to be told. That’s worth knowing.
3. “Where exactly did you get stuck?”
Not did you get stuck. Where. The word “exactly” does the work: it pushes from “I didn’t get it” to “I didn’t know what to do after I found x”. The first is a mood. The second is a fixable gap, and it’s usually one line long.
4. “What would you do differently?”
Asked after it’s finished, right or wrong. This is the one that transfers to next time. Without it, a corrected mistake stays a corrected mistake. With it, it becomes a rule they wrote themselves.
How to ask them without it becoming an interrogation
Three things make the difference between this working and this being annoying.
Ask them when things went right, not only when they went wrong. If the four questions only appear after a bad mark, they become a punishment and the child learns to give you the shortest possible answers. Asking “how sure were you?” after a correct answer is where most of the value is anyway.
Accept “I don’t know” and wait. The pause is the work. Children aren’t used to being asked these and the first few answers will be thin. That’s fine.
Don’t follow up with the answer. The moment you say “well, you should have used the formula”, the questions stop being about their thinking and start being a delivery mechanism for yours. Ask, listen, stop.
This connects to something we’ve written about before: a child explaining their reasoning out loud learns more than a child being explained to. These four questions are the smallest possible version of that, and they fit into the gap between finishing homework and putting it away.
Frequently asked questions
What is metacognition in simple terms?
Thinking about your own thinking - knowing what you understand, noticing when you’re stuck, and being able to describe what you tried.
How big is the effect?
The Education Endowment Foundation rates metacognition and self-regulation at +8 months of additional progress per year, based on 355 studies. The figure was revised upwards from +7 when the review was updated in May 2025.
What age does this work from?
EEF finds benefits across primary and secondary. Younger children need the questions asked out loud by an adult; by secondary, the goal is that they ask themselves.
My child just says “I don’t know.”
Normal at the start. Wait rather than filling the silence, and ask about successes as well as failures so the questions aren’t associated with being in trouble.
Isn’t this just asking them to explain their working?
Related but not the same. Explaining working is about the maths. These questions are about the child’s awareness - confidence, strategy choice, where the breakdown was. A child can show perfect working and still have no idea whether they trust it.
Does this replace actually practising?
No. It makes practice count for more. Pair it with mixed practice, where the questions are shuffled across chapters - the confidence question gets far more interesting when the child doesn’t know what type of problem is coming.
Related reading
- How to Study Maths Effectively - five shuffled questions - the practice format these questions work best with.
- Your Child Can Get Every Answer Right and Still Not Understand - why a correct answer is weak evidence on its own.