Interactive Teaching and Gamification: What the Evidence Says About the Impact on Children
Games, points and interactive apps are now part of many children's learning. The research says they can help, but by less than the marketing suggests, and the design choices decide whether they help, do nothing or get in the way.
Why this matters
Schools, parents and funders are asked to choose between hundreds of products that promise engagement. Engagement is not the same as learning. This report brings together the strongest syntheses and UK trials we could verify, so that decisions about children's learning technology rest on evidence rather than on how engaging a demo looks.
Key findings
- Effects are real but modest, and they shrink under stricter evaluation. Digital games beat ordinary teaching by g = 0.33 across all studies, but by 0.17 in randomised studies alone.[6]
- How learning is measured changes the answer. Apps for children aged about 3 to 9 show predicted effects of 0.43 on tests researchers designed, but 0.17 on standardised tests.[3]
- Most gamification research is not about children. In the most-cited meta-analysis, only 3 of 19 learning-outcome studies took place in schools.[11]
- Large UK trials show modest results. Estimates range from three months of additional progress to minus one month, including a reading game that pupils and staff found highly engaging.[25][26]
- Design decides. Game elements that carry the learning content outperform elements added for fun (g = 0.43 against 0.18).[9]
- Some mechanics backfire with children. Expected tangible rewards reduced children's later interest in a task more than older students' (d = −0.39 against −0.27), and 80% of apps used by 3 to 5-year-olds contained manipulative design.[16][21]
Figure 1
Stricter evaluation, smaller effects
The same body of evidence, read two ways: a less strict view and a stricter one.
-
Digital games against ordinary teaching Clark et al. 2016 · ages 6 to 25 · all studies, then randomised studies only0.17 0.33
-
Learning apps, ages 3 to 9 Kim et al. 2021 · predicted effect on researcher-designed tests, then standardised tests0.17 0.43
-
Learning apps, ages 3 to 9 (2026 update) Gilbert et al. 2026 · modelled effect: small quasi-experiment on a researcher test of a narrow skill, then a large randomised trial on a standardised test of a broad skill0.07 0.94
Effect size (standardised mean difference; 0.2 is conventionally small, 0.5 medium)
Show Figure 1 as a table
| Evidence | Less strict view | Stricter view |
|---|---|---|
| Digital games against ordinary teaching (Clark et al. 2016) | All studies: g = 0.33 | Randomised only: g = 0.17 |
| Learning apps, ages 3 to 9 (Kim et al. 2021) | Researcher-designed tests: 0.43 | Standardised tests: 0.17 |
| Learning apps, ages 3 to 9 (Gilbert et al. 2026, modelled) | Small quasi-experiment: about 0.94 | Large randomised trial: about 0.07 |
What this report covers, and how we compiled it
Three ideas are often blurred together, and the evidence differs for each:
- Interactive teaching asks learners to do something with the material: answer, explain, build or discuss.
- Game-based learning uses a game as the learning activity itself.
- Gamification adds game elements, such as points, badges, levels, leaderboards and stories, to an activity that is not a game.
We reviewed meta-analyses, systematic reviews and large trials published up to September 2026, giving priority to studies with children aged 3 to 16, together with the UK guidance and regulation that applies to products children use. Where evidence comes from older learners, we say so. We report effect sizes as the papers report them (Hedges' g or Cohen's d), and UK trial results in the Education Endowment Foundation's months of additional progress. This is a structured review of existing syntheses, not a new meta-analysis: we have not pooled studies ourselves.
Interactive is not the same as tapping
Chi and Wylie's ICAP framework sorts learning activities by what learners visibly do: passive (receiving), active (manipulating), constructive (producing something beyond what they were given) and interactive (building on each other's contributions in dialogue). It predicts that learning improves in that order.[1] The authors point out that software is often called interactive when it is not: choosing an answer from a menu is only active, and a system counts as interactive only when it responds to the substance of what the learner says or does.
For young children, Hirsh-Pasek and colleagues set out four pillars for apps that genuinely teach. Children should be actively thinking rather than just tapping, engaged with the learning goal rather than distracted from it, working with meaningful content, and interacting socially with adults or other children.[2] They warn that pop-up features which interrupt a story can harm young children's comprehension, and that informative feedback serves children better than stickers and praise.
Apps for young children: real gains, smaller on independent tests
The strongest synthesis for young children is Kim and colleagues' meta-analysis of 36 studies of apps for children from preschool to US grade 3, roughly ages 3 to 9. Apps improved early literacy and maths, with an average effect of g = 0.31. The size of the effect depended on how learning was measured: the model predicted 0.43 on tests the researchers designed and 0.17 on standardised tests, and larger effects on narrow skills such as letter names than on broad ones such as vocabulary.[3] Time spent on the apps did not consistently predict the size of the effect.
A 2026 update covering 141 studies confirms the pattern. Randomised trials reported smaller effects than other designs, and larger studies smaller effects than small ones. The model implies an effect of about 0.07 for a large randomised trial measured on a standardised test of a broad skill, against about 0.94 for a small quasi-experiment measured on a researcher-made test of a narrow skill (Figure 1).[4] That paper is a working paper and not yet peer reviewed. A systematic review in Pediatrics found learning benefits for children under 6, especially in early maths, but the studies varied widely and were often poorly reported.[5]
Game-based learning at school
Clark and colleagues' meta-analysis of 69 study samples (6,868 participants, average age about 13) found that digital games beat ordinary teaching by g = 0.33. Two findings matter more than the average. Randomised studies found about half that effect (0.17), and games played over several sessions did far better than one-off play (0.44 against 0.08). Enhanced scaffolding, meaning guidance that helps learners make sense of what happens in the game, added g = 0.41 over the same game without it.[6]
Wouters and colleagues reached similar conclusions across 39 studies. Learning improved (d = 0.29) and retention more so (d = 0.36), but games were not significantly more motivating than conventional teaching, and in randomised studies games were not more effective at all. Games worked best combined with other teaching, over several sessions, and played in groups (d = 0.66, against 0.22 when played alone).[7]
The strongest recent synthesis restricted to school pupils found an overall effect of g = 0.54: 0.67 on knowledge and skills, 0.32 on motivation and attitudes, and no significant effect on metacognition.[8] A 2023 meta-analysis of STEM games, whose samples were about three-quarters school-age, identified the design lesson that runs through this whole literature: game elements that carry the learning content produced g = 0.43, against 0.18 for elements added for the gaming experience, such as points or leaderboards.[9] In maths alone, a meta-analysis of 24 studies from preschool to the end of secondary school found a small effect of d = 0.13.[10]
Gamification: points, badges and leaderboards
The most-cited meta-analysis of gamification in education found small positive effects: g = 0.49 on learning, 0.36 on motivation and 0.25 on behaviour, with only the learning effect holding up in the most rigorous studies. For anyone designing for children, the crucial detail is who was studied: only 3 of its 19 learning studies and 3 of its 16 motivation studies took place in schools. Most involved university students.[11]
Two of its moderator findings are especially useful. Competition on its own did not significantly improve motivation or behaviour, while competition combined with collaboration did (g = 0.63 for motivation and 0.70 for behaviour). And longer interventions produced larger motivational effects than one-day ones.[11]
Other syntheses add nuance. One found g = 0.50 across 24 studies, while qualitative studies in the same review showed that some learners found gamification pointless, or a source of anxiety or jealousy.[12] Another found only a small effect on intrinsic motivation (g = 0.26) and a minimal effect on learners' sense of competence.[13]
Newer syntheses restricted to school pupils report larger pooled effects: g = 0.65 on motivation across K-12 studies, with very high variation between studies,[14] and g = 0.82 on knowledge in K-12 STEM subjects.[15] Yet the effect on primary pupils' motivation was small (g = 0.31),[14] and we found no large, independent randomised trial that isolates points, badges or leaderboards with school pupils. The honest summary: gamification can help, most of the evidence comes from older learners, and the larger effects reported for school pupils vary widely and have not yet been tested at scale.
What happens at scale: UK trials
Meta-analyses average many small studies. Large trials in real schools show what happens when a programme is used at scale. The Education Endowment Foundation (EEF) has tested several that use games, apps or game mechanics.
| Programme | What it is | Scale | Result |
|---|---|---|---|
| Mathematical Reasoning[24] | Year 2 maths lessons that include computer games | 238 schools, 6,443 pupils | +1 month; +2 for pupils eligible for free school meals |
| onebillion maths apps[25] | Year 1, lower-attaining half, supervised by teaching assistants | 113 schools, 1,124 pupils | +3 months; −2 for pupils eligible for free school meals (less secure) |
| GraphoGame Rime[26] | Year 2 phonics game for pupils behind in reading | 15 schools, 398 pupils | −1 month, although pupils and staff found it highly engaging |
| Good Behaviour Game[27] | Key Stage 2 team-points game for classroom behaviour | 77 schools, 3,084 pupils | 0 months on reading, no effect on behaviour; delivered with low fidelity |
Two smaller studies help explain the spread. In a randomised trial with 389 children aged 4 to 5, maths apps raised attainment more when they supplemented teaching (0.31) than when they replaced it (0.21).[28] And an international review of GraphoGame found no overall effect on word reading (g = −0.02), but an effect of 0.48 where adults interacted with children during play.[29] Engagement, in other words, is not evidence of learning. How the technology sits inside teaching matters more.
Where effects fade or backfire
Rewards that crowd out interest
A meta-analysis of 128 experiments found that expected tangible rewards for a task people already found interesting reduced their later, voluntary engagement with it (d = −0.36). The effect was stronger for children than for college students (−0.39 against −0.27). Unexpected rewards had no such effect, and positive feedback, which raised college students' motivation, did not significantly raise children's (d = 0.11).[16] These were short laboratory tasks rather than digital products, but the implication for designers is direct: paying children in points or prizes for doing something they enjoy can teach them to stop when the payments stop.
Novelty that wears off
In a semester-long university course, a section gamified with a leaderboard and badges ended with lower motivation, satisfaction and exam scores than an identical course without them.[17] A 14-week study of 756 undergraduates found that gamification's effect dipped after about four weeks and then partly recovered.[18] Long-term evidence with children is thin, but the meta-analyses agree that repeated sessions inside teaching beat one-off use, and that more time on an app does not by itself mean more learning.[6][7][3]
Competition and public rankings
Single-player competitive games showed no benefit over ordinary teaching (g = −0.06, from only four studies), against 0.45 for non-competitive single-player games.[6] Competition alone did not improve motivation,[11] and learners in gamified classes reported anxiety and jealousy.[12] Team challenges have better evidence than individual leaderboards.
Distracting detail
Interesting but irrelevant additions to learning material, known as seductive details, reduce learning.[19] A 2026 meta-analysis of 50 studies put the overall cost at g = −0.16, working mainly through the attention they consume.[20] In games, simple schematic visuals outperformed realistic ones.[6][7]
Manipulative design
A study of 160 children aged 3 to 5 in the United States examined the 133 apps they played longest. Only 20% contained no manipulative design. Features that pressured children to keep playing appeared in 64.7% of apps, pressure to buy in 55.6% and advertising in 31.6%. Free apps contained more manipulative features than paid ones, and children whose parents had less education were exposed to more of them.[21]
Screen time for the youngest
The World Health Organization recommends no sedentary screen time for children under two, and no more than one hour a day for children aged 2 to 4, with less being better.[22] Its guidance makes no exception for educational content.
The UK rules that now apply
Teaching guidance. The EEF's guidance on digital technology recommends deciding how technology will improve teaching before introducing it, and using it to improve explanations, pupil practice, and assessment and feedback. It stresses that pedagogy rather than the technology drives impact, and that technology works best supplementing teaching.[23] In the EEF's toolkit, feedback is one of the highest-impact approaches, at about six months of additional progress.[33]
Data protection. The ICO's Children's code has applied since September 2021. Its guidance on the detrimental use of data names reward loops, continuous scrolling, notifications and autoplay as features that can keep children engaged against their interests, and asks services to let children pause without losing progress and to present options to keep playing neutrally. Privacy settings should be high by default and profiling off by default. The code applies to education technology sold directly to families, and to school-provided products where the provider decides how data is used beyond the core service.[30] The ICO's June 2026 review of 28 edtech providers, used by more than 90% of UK schools, made 596 recommendations. Common problems included providers misjudging whether they acted as controller or processor, and missing or thin data protection impact assessments.[31]
AI products. The Department for Education's product safety standards for generative AI in education, updated in January 2026, rule out manipulative strategies such as flattery, pressure from comparison with peers, guilt and threats of loss. Rewards are acceptable only when transparent, low-stakes and educationally justified; products should not artificially prolong use; learners should get progressive hints rather than full answers by default; and claims of impact must be backed by robust evidence.[32]
Eight design principles for children's interactive learning
- Make the learning the game. The core action should require the target thinking, such as predicting, explaining or building, not just tapping. Add game elements that carry the content.[9][1][2]
- Put informative feedback and scaffolding ahead of points. Tell children why an answer is right or wrong and what to try next; praise alone did not lift children's motivation.[6][16][33]
- Don't pay children to play. Avoid expected rewards for taking part in or finishing activities that are already interesting. Keep any rewards symbolic, transparent and low-stakes.[16][32]
- Design for talk and teamwork. Build in adult and peer interaction, and prefer team challenges to individual public rankings.[29][7][11]
- Remove seductive detail. Use simple visuals and stories only where they serve the learning.[19][20]
- No dark patterns. Let children stop without losing progress; avoid streak and loss pressure, lures, adverts and purchase prompts; keep privacy high by default.[21][30][32]
- Build for sustained use inside teaching. Plan for several sessions integrated with lessons, and expect novelty to fade.[6][23][28]
- Measure learning independently. Use standardised or independent measures, report engagement separately, and check who benefits.[3][4][26]
What this means for schools, researchers and product teams
For schools, the question to ask of any product is not whether children enjoy it but what independent evidence shows it improves, for whom, and how it should sit inside teaching. Engagement features deserve the same scrutiny as learning claims, including against the Children's code.
For researchers, the evidence rewards rigour: independent measures, randomised designs where possible, interventions long enough for novelty to fade, and results reported for disadvantaged pupils as well as on average.
For product teams, the design lesson is consistent. The mechanics that help children learn are the ones that carry the learning: feedback, scaffolding, meaningful challenge and interaction with other people. The mechanics that only hold attention, such as points for their own sake, public rankings and pressure to keep playing, have weak or negative evidence, and are increasingly regulated.
How we apply this at EdTechLab
Most of the platforms we run serve older learners, but the same evidence shapes them. EngagedLab's Socratic tutor withholds answers while a learner is still attempting a task. intle's scenario questions explain why each option is right or wrong rather than only scoring it. WorkReady Finance, which we built for UWE Bristol, puts decisions with consequences at the centre, with short explanations beside each one, and learners stay anonymous. Any product we build for children will be designed to these principles and to the Children's code from the first sketch.
Limits of this report
- Effect sizes from different syntheses use different comparisons and measures, so they are not directly comparable.
- Many underlying studies are small, short and use researcher-designed tests, which tend to inflate effects.
- Much of the gamification evidence comes from university students rather than children.
- One source is a working paper not yet peer reviewed, and for several recent K-12 syntheses we relied on the published abstract.
- We did not assess the risk of bias of individual studies or pool results ourselves.
References
- Chi MTH, Wylie R. The ICAP framework: Linking cognitive engagement to active learning outcomes. Educational Psychologist, 2014, 49(4), 219–243. DOI
- Hirsh-Pasek K, Zosh JM, Golinkoff RM, et al. Putting education in "educational" apps: Lessons from the science of learning. Psychological Science in the Public Interest, 2015, 16(1), 3–34. DOI
- Kim J, Gilbert J, Yu Q, Gale C. Measures matter: A meta-analysis of the effects of educational apps on preschool to grade 3 children's literacy and math skills. AERA Open, 2021, 7. DOI
- Gilbert JB, Young WS, Kim JS. Measures still matter: Replicating and extending a meta-analysis of the effects of educational apps on children's reading and math skills with six years of new data. EdWorkingPaper 26-1579, 2026 (not yet peer reviewed). DOI
- Griffith SF, Hagan MB, Heymann P, Heflin BH, Bagner DM. Apps as learning tools: A systematic review. Pediatrics, 2020, 145(1), e20191579. DOI
- Clark DB, Tanner-Smith EE, Killingsworth SS. Digital games, design, and learning: A systematic review and meta-analysis. Review of Educational Research, 2016, 86(1), 79–122. DOI
- Wouters P, van Nimwegen C, van Oostendorp H, van der Spek ED. A meta-analysis of the cognitive and motivational effects of serious games. Journal of Educational Psychology, 2013, 105(2), 249–265. DOI
- Barz N, Benick M, Dörrenbächer-Ulrich L, Perels F. The effect of digital game-based learning interventions on cognitive, metacognitive, and affective-motivational learning outcomes in school: A meta-analysis. Review of Educational Research, 2024, 94(2), 193–227. DOI
- Gui Y, Cai Z, Yang Y, Kong L, Fan X, Tai RH. Effectiveness of digital educational game and game design in STEM learning: A meta-analytic review. International Journal of STEM Education, 2023, 10, 36. DOI
- Tokac U, Novak E, Thompson CG. Effects of game-based learning on students' mathematics achievement: A meta-analysis. Journal of Computer Assisted Learning, 2019, 35(3), 407–420. DOI
- Sailer M, Homner L. The gamification of learning: A meta-analysis. Educational Psychology Review, 2020, 32(1), 77–112. DOI
- Bai S, Hew KF, Huang B. Does gamification improve student learning outcome? Evidence from a meta-analysis and synthesis of qualitative data in educational contexts. Educational Research Review, 2020, 30, 100322. DOI
- Li L, Hew KF, Du J. Gamification enhances student intrinsic motivation, perceptions of autonomy and relatedness, but minimal impact on competency: a meta-analysis and systematic review. Educational Technology Research and Development, 2024, 72(2), 765–796. DOI
- Kurnaz MF, Koçtürk N. A meta-analysis of gamification's impact on student motivation in K-12 education. Psychology in the Schools, 2025, 62(12), 4997–5009. DOI
- Cheng M, Lai X, Zeng W, Bai S, Duan C, Sun D. Unlocking the power of gamification in K-12 STEM education: A meta-analysis of the effects on domain knowledge, higher-order thinking, and affective outcomes. Educational Psychology Review, 2026, 38, 60. DOI
- Deci EL, Koestner R, Ryan RM. A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 1999, 125(6), 627–668. DOI
- Hanus MD, Fox J. Assessing the effects of gamification in the classroom: A longitudinal study on intrinsic motivation, social comparison, satisfaction, effort, and academic performance. Computers & Education, 2015, 80, 152–161. DOI
- Rodrigues L, Pereira FD, Toda AM, et al. Gamification suffers from the novelty effect but benefits from the familiarization effect: Findings from a longitudinal study. International Journal of Educational Technology in Higher Education, 2022, 19, 13. DOI
- Sundararajan N, Adesope O. Keep it coherent: A meta-analysis of the seductive details effect. Educational Psychology Review, 2020, 32(3), 707–734. DOI
- Cheng C, Wu Y, Wang R, Wang Z. Seductive details, cognitive load, and learning outcomes: A multi-level meta-analysis and MASEM. Educational Psychology Review, 2026, 38, 28. DOI
- Radesky J, Hiniker A, McLaren C, et al. Prevalence and characteristics of manipulative design in mobile applications used by children. JAMA Network Open, 2022, 5(6), e2217641. DOI
- World Health Organization. Guidelines on physical activity, sedentary behaviour and sleep for children under 5 years of age. 2019. Source
- Stringer E, Lewin C, Coleman R. Using digital technology to improve learning: Guidance report. Education Endowment Foundation, 2019. Source
- Education Endowment Foundation. Mathematical Reasoning, 2023–24 effectiveness trial. Source
- Education Endowment Foundation. onebillion: Year 1 pupils learning maths on apps. Source
- Education Endowment Foundation. GraphoGame Rime. Source
- Education Endowment Foundation. The Good Behaviour Game. Source
- Outhwaite LA, Faulder M, Gulliford A, Pitchford NJ. Raising early achievement in math with interactive apps: A randomized control trial. Journal of Educational Psychology, 2019, 111(2), 284–298. DOI
- McTigue EM, Solheim OJ, Zimmer WK, Uppstad PH. Critically reviewing GraphoGame across the world: Recommendations and cautions for research and implementation of computer-assisted instruction for word-reading acquisition. Reading Research Quarterly, 2020, 55(1), 45–73. DOI
- Information Commissioner's Office. Age appropriate design code (Children's code) guidance and resources, including edtech guidance. Source
- Information Commissioner's Office. Edtech examined, June 2026. Source
- Department for Education. Generative AI: product safety standards, updated January 2026. Source
- Education Endowment Foundation. Teaching and Learning Toolkit: Feedback. Source
Continue the reading
Designing or evaluating learning for children?
We design, build and evaluate learning platforms with the evidence in this report built in, from the first sketch to independent measures of what changed.