The Evidence on Teaching

The Impact of an Acute Bout of Physical Activity on Executive Function and Academic Achievement in Preadolescent Children: A Systematic Review and Meta-Analysis

Dixon S, Tempest G, Hettinga F, Khudair M, McCullogh N · 2025

grade Cmeta-analysisindependentfailed
Sample
15 studies; inhibition (n=13), working memory (n=9), cognitive flexibility (n=7); only 2 studies measured academic achievement at all
Population
Preadolescent children
Design
PRISMA systematic review and meta-analysis restricted to acute (single-bout) physical activity in preadolescents — the correct age scope for the classroom-activity-break question, unlike Chang et al. 2012's lifespan corpus. Authors report high heterogeneity, moderate risk of bias, and detected publication bias. Graded C on corpus quality and size.
Key findings
The acute effect that classroom activity breaks are sold on is not there in the target population. Pooled effect on executive function is d=0.02 for accuracy (95% CI -0.04 to 0.07) and d=-0.02 for response time (95% CI -0.40 to 0.36) — both centred on zero, with the response-time interval spanning a wide range in both directions. Publication bias was detected even in this small corpus. And the fact that matters most for schools: only 2 of 15 studies measured academic achievement at all, too few to pool. The acute-exercise literature has essentially never tested the outcome schools care about.
Genetic confound
Acute crossover and randomized-order designs are not genetically confounded. The genetic question does not arise here; the problem is that the effect is absent, not that it is confounded.

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Executive function, accuracyCohen's d0.02 (95% CI -0.04 to 0.07)standardizedacute; post-bout; lab tasksbusiness-as-usualend-of-treatmentnear-transfer
Executive function, response timeCohen's d-0.02 (95% CI -0.40 to 0.36)standardizedacute; post-bout; lab tasksbusiness-as-usualend-of-treatmentnear-transfer
Academic achievementnumber of studiesonly 2 of 15 studies measured it — meta-analysis not possibleunknownacutebusiness-as-usualend-of-treatmentdomain-skill
Publication biasqualitativedetected; heterogeneity high; risk of bias moderatestandardizedacross corpusnonenot-applicablenear-transfer

Cited by