Physical activity linked to small gains in children's attention and math performance

In the primary analysis, physical activity showed small pooled benefits for attention (standardized mean difference [SMD] = 0.19) and mathematics (SMD = 0.21). However, trim-and-fill analyses—which account for potential missing small studies—indicated that adjusted confidence intervals included zero, suggesting the pooled estimates may be vulnerable to publication bias and small-study effects. When the researchers examined which intervention types worked best, patterns emerged but remained descriptive rather than statistically tested. For attention, longer individual sessions (30–45 minutes) appeared more favorable (SMD = 0.30) than shorter ones, fewer sessions per week (≤3) showed slightly larger effects, and shorter programs (< 6 weeks) had the largest descriptive effect (SMD = 0.42). For mathematics, the pattern differed: shorter sessions (< 30 minutes) looked most promising (SMD = 0.30), and physically active academic lessons—lessons that integrated mathematics directly with movement—showed encouraging trends. The authors emphasize that these subgroup findings should be interpreted cautiously. Formal statistical tests comparing subgroups were not performed, heterogeneity across studies was substantial, and some network estimates were imprecise. The small effect sizes observed mean that while physical activity may support both attention and math learning, the benefit per intervention is modest in magnitude.

The meta-analysis found that for attention, very short intervention periods (< 6 weeks) had the largest descriptive effect (SMD = 0.42), suggesting rapid cognitive gains are possible—but this pattern came from few trials and needs replication. Resistance training ranked highest when interventions were ranked by likelihood of benefit (SUCRA analysis), hinting that structured strength work may have particular value for attention in children. Conversely, mathematics showed its largest descriptive effect with 6–9 week programs, a longer window than optimal for attention, implying that math achievement may require more sustained behavioral change to register.

Critical limitations: the researchers did not formally test whether attention and math truly benefit from different exercise parameters—these were observed patterns in subgroups, not pre-specified hypotheses. Publication bias is a real concern; when the team imputed studies likely missing from published literature, the confidence intervals shifted to include zero, meaning the true effect could be smaller or even absent. Effect sizes of 0.19–0.21 are considered small in educational research, roughly equivalent to 2–3 months of additional learning gain. The heterogeneity across studies was high, meaning variation between individual trials was large and may reflect differences in how physical activity was delivered, children's baseline fitness, or measurement tools used.