Play-based maths practice for ages 5 to 12: what the research supports, and what to avoid

The short answer

1. Motivation goes up. A 2024 systematic review by Ramírez Ruiz, Vargas Sanchez and Boude Figueredo, covering 90 school studies, found children took part more and were more motivated when lessons added points, levels and challenges. The same review says that did not show up as significant gains in achievement against ordinary teaching, and that most studies were too short and small to tell.
2. The maths-specific meta-analysis (Tokac, Novak and Thompson, 2019, 24 studies) found a small average benefit from learning maths through digital play, with wide variation between products.
3. Short trials look best. Wang and colleagues (2022, 33 STEM studies) found the largest effects in primary school and in interventions under a week, which the authors themselves put down to novelty.
4. Expected prizes cost something. Deci, Koestner and Ryan (1999, 128 experiments) found that a prize a child expects for doing a task, certificates included, lowers the child's own motivation once the prize stops, and more so in children than in adults.
5. What the Education Endowment Foundation backs is practice: more of it, spaced, retrieval-based, low stakes, with fast recall of number facts as a foundation.

What the reviews found

That last row is the only study of a PYP classroom found, and it is about Mandarin, not maths. There is no study of the programme's maths phases and play-based practice. Anyone who tells you there is should show it to you.

Rewards and young children

The rewards evidence is older and stronger than the digital evidence. Deci, Koestner and Ryan pooled 128 experiments in 1999: when a child expects a tangible reward for doing a task, their free-choice motivation afterwards drops, and a "good player" certificate does it as surely as sweets. The drop was larger for children than for students. Praise helped students; in children the effect was small and not significant. Unexpected rewards, and rewards unrelated to the task, did nothing either way.

Those were experiments, not classrooms, and nobody in that literature tested digital badges or class rankings on a screen with five to twelve year olds. So the honest position is: expected prizes carry a known cost with young children; public rankings are untested and we avoid them as a precaution, not because a study condemned them.

What the Education Endowment Foundation says

• Individualised instruction (Toolkit, 198 studies): about three months' extra progress in primary and in maths; digital versions do about as well as paper; a supplement to teaching, not a replacement.
• Using digital technology to improve learning (guidance, 2019): technology helps most when it gives more practice of better quality; retrieval, spaced practice and low-stakes quizzing improve retention; motivation may not turn into learning; adaptive programmes have on average done less well than simpler extra-practice ones.
• Improving mathematics in Key Stages 2 and 3: quick recall of number facts matters; technology is "not a silver bullet".
• Improving mathematics in the early years and Key Stage 1 (2020): play gives repeated practice in a motivating context; linear numbered board games help number understanding; a computer does not make a programme effective on its own.

The EEF does not name or rate any product on this page, and there is no longer a Toolkit strand for digital technology as such.

What that means for a class

Do: short and often (five to ten minutes, most days) rather than one long session; space the practice, so the hard facts come back after a day and again after a week; keep it low stakes, with no mark that follows the child; put the practice on the facts that child got wrong, not the ones they enjoy; keep the teacher conversation for the unit and the tool for the facts.

Avoid: prizes a child expects for finishing; rankings the class can see; long screen sessions; judging any tool by its first week, which is the novelty week; and any product that promises months of progress.

How Explorisle applies this, and what it cannot claim

Explorisle is built on the practice side of this evidence: a question at every post at the child's level, the wrong ones asked again later and again next week, a way to see it beside every wrong answer, a stop offered at about five minutes. There are no prizes for finishing, no streak, no rank, and no score that leaves the child's device. What the child earns is planks and gems to build with, tied to the task itself.

There is no study of Explorisle. As of today it has been played by a handful of children in one class test. Nothing on this page is a claim about what it does to marks, and if a study is ever run, the result will be published here whichever way it goes.

Sources

• Ramírez Ruiz, Vargas Sanchez and Boude Figueredo (2024), systematic review, open access: the review
• Tokac, Novak and Thompson (2019), Journal of Computer Assisted Learning 35(3): ERIC record
• Wang, Chen, Hwang, Guan and Wang (2022), International Journal of STEM Education 9:26: full text
• Sailer and Homner (2020), Educational Psychology Review 32(1): ERIC record
• Wijaya and Lukas (2025), EDUTECH 24(3): abstract
• Deci, Koestner and Ryan (1999), Psychological Bulletin 125(6), restated in Review of Educational Research 71(1), 2001
• Education Endowment Foundation: Individualised instruction, Using digital technology to improve learning, Improving mathematics in Key Stages 2 and 3, Improving mathematics in the early years and Key Stage 1

Read on 26 September 2026. Where a full text was behind a paywall, the record read is linked instead.