Mathematics

How to Build Maths Confidence

Confidence is a memory of succeeding at something hard, not a feeling you can talk a child into. Aim for four questions in five right, and praise the method.

Short answer

How do you build confidence in maths?

By changing the work, not the words. Set questions the child gets right about four times in five, and praise the method they used rather than how clever they are.

The short answer
Confidence in maths is a memory of having succeeded at something that was hard, so it is built by changing what a child works on rather than what you say about them. Aim for work they get right roughly four times in five, and praise the method and the persistence rather than the ability.

What changes the answer

A child whose confidence collapsed after one specific event — a test, a set move, a remark — needs that event named and addressed, not a general programme of gentler work.
Below about age eight the fastest lever is usually arithmetic fluency, because a child still counting on their fingers has no attention left over to feel capable with.
Where the reaction is physical — tears, nausea, refusal to enter the room — this is past difficulty tuning and the school's pastoral team should be involved.
A child who is genuinely under-challenged looks unconfident too, because nothing they do feels like an achievement; for them the answer is harder work, not kinder work.

Why praise aimed at ability backfires

Mueller and Dweck ran six studies with fifth-graders in 1998. Children praised for intelligence after a success, compared with children praised for effort, went on to choose easier tasks, persisted less after a failure, enjoyed the work less, performed worse, and were more likely to describe ability as a fixed trait.

The mechanism is easy to feel once it is named. 'You're so clever at maths' makes every subsequent difficulty evidence against the claim, and the only move available to protect the claim is to avoid the difficulty. 'You worked through that without giving up' names something the child chose and can choose again tomorrow.

This does not mean praise nothing. It means praise the part that was a decision: the method, the persistence, the choice to check, the willingness to start a question they were not sure about.

The difficulty band nobody adjusts

Work a child gets right every time teaches nothing, and the child knows it — praise for it is heard as pity. Work they get right four times in ten is simply humiliating, and no amount of encouragement makes an hour of it feel good. The productive band sits high but not at the top.

Wilson and colleagues derived an optimal training accuracy of about 85% for a broad class of learning algorithms. That is a result about algorithms rather than about children, and it should be treated as a target rather than a law — but it lines up closely with the level experienced teachers set by instinct, and it gives a parent a number to aim at.

In practice: set ten questions and expect eight right. Ten out of ten means the set was too easy and the next one should be harder. Five out of ten means back off. This is a dial the adult controls completely, and it is the one almost nobody turns.

Make the child's own progress visible to them

Confidence is built from evidence, and the evidence is usually thrown away. Loose worksheets go in the bin; the child has no record that September ever happened, so every session starts from a standing position.

Keep one book rather than loose sheets, so that September is still in it in March.
Once a month, set a question from two months ago and let them notice for themselves.
Name the specific change: 'in March you could not rearrange that. You just did it without asking.'
Say 'you have not learnt that yet' rather than 'you are not a maths person'.
Stop rescuing at the first sign of struggle. A parent who steps in immediately teaches that struggle is a signal to stop.

Why a small group does something one-to-one cannot

A child alone with a tutor never hears anybody else get anything wrong. Every error in the room is theirs, and after a term of that the arithmetic is unavoidable: they are the person in this room who makes mistakes.

In a group of four they hear it several times an hour. Discovering that the confident-looking classmate also loses the minus sign is worth more than any amount of reassurance from an adult, because it is not an opinion — it is something they watched happen. That is a real argument for small groups and we make it deliberately rather than as a consequence of pricing.

Training accuracy identified as optimal for a broad class of learning algorithms: 85%Training accuracy identified as optimal for a broad class of learning algorithms[2]

Common questions

Is it confidence or ability that is the problem?

Test it. Give the same questions with no marks attached, no time limit and nobody watching. A child who can do them alone but not in class has a confidence problem. A child who cannot do them either way has a knowledge problem, and reassurance will not touch it — which is worth finding out before spending a term on the wrong one.

Should I tell my child I was bad at maths too?

No. It is offered as solidarity and it lands as permission — and often as a family diagnosis. If you want to say something about your own experience, say that you found it hard and that nobody showed you how, which locates the problem in the teaching rather than in the bloodline.

How long does building confidence take?

Weeks before you see a child attempt something they would previously have skipped, and a term or more before it looks like confidence from the outside. The first observable change is almost always the willingness to start a question they are unsure of, and it is worth watching for, because it arrives well before the marks do.

Does praise ever help?

Yes, when it names something the child chose: the method, the persistence, the decision to check the answer. What does not help is praise for being clever, praise for speed, or praise for work that was easy. Children detect the last one instantly, and it devalues everything else you say afterwards.

Sources

Praise for intelligence can undermine children's motivation and performance — Journal of Personality and Social Psychology (Mueller & Dweck, 1998)
The Eighty Five Percent Rule for optimal learning — Nature Communications (Wilson, Shenhav, Straccia & Cohen, 2019)

Last updated August 21, 2026