Gatiप्रवाह

The evidence

WHAT THE
RESEARCH
ACTUALLY SAYS

Every study on this page is checkable and none of them are ours. Each card carries what the paper found, the numbers it reported drawn to scale, and the thing it does not show. That last part usually goes missing, so here it is set in the same size as the rest.

How to read the numbers

THE LINE AT ZERO
DECIDES IT

An established finding

The whole range sits clear of zero, so the direction of the effect is settled by it.

A finding that is not established

The range includes zero, so the data still allows for no effect at all.

The dot
The single best estimate the study arrived at.
The bar
The 95% confidence interval: the range the true effect most likely sits in. A wide bar is an uncertain size.
The line at zero
No effect. A bar that touches it has not established one, however good the headline number looks.

Effect sizes are measured in standard deviations. By convention 0.2 is small, 0.5 moderate and 0.8 large, and those three marks are ruled onto every chart below. The two rows above are an illustration of the two states; every figure after this point is measured and cited.

01

Movement and learning are related, but the link is not automatic

64 studies, children aged 5–13

The College reviewed everything published since 1990 on physical activity, fitness, cognition and attainment in children. It found consistent positive associations, and evidence that activity engages the brain regions that carry complex cognitive work.

What the panel concluded

“promising but uncertain”

The strongest claim the panel was willing to make, after reviewing everything published on the question since 1990.

What it does not show: The panel called the case for using activity to raise attainment "promising but uncertain". An association still has to be turned into a lesson plan that works, and this is the paper most often quoted as though it had done that.

Donnelly, Hillman, Castelli, Etnier, Lee, Tomporowski, Lambourne & Szabo-Reed (2016). Medicine & Science in Sports & Exercise, 48(6), 1197–1222. American College of Sports Medicine position stand. Read the record

02

Physical education moves classroom behaviour more than it moves marks

26 intervention studies, 10,205 children aged 4–13

A meta-analysis of activity interventions against academic outcomes. The largest single effect was not on any subject. It was on whether children stayed on task once they were back at their desks.

On-task classroom behaviour

0.77 (0.22 to 1.32)

Composite academic score

0.26 (0.07 to 0.45)

Mathematics

0.21 (0.09 to 0.33)

Reading

0.13 (0.02 to 0.24)

Language skills

0.16 (−0.06 to 0.37)

Not established

00.20.50.8

standard deviations

One interval here reaches the zero line. That finding is unestablished, and it is drawn hollow and dashed to keep it from being read as a result.

What it does not show: The language-skills interval crosses zero, so that one is not established. The subject-level effects that do hold are small. The behaviour effect is large but has a wide interval, which means the size is uncertain even though the direction is not.

Álvarez-Bueno, Pesce, Cavero-Redondo, Sánchez-López, Garrido-Miguel & Martínez-Vizcaíno (2017). Pediatrics, 140(6), e20171498. Read the record

03

How a lesson is designed changes how much of it is actually active

42 studies, 12,663 participants, 27 randomised trials

The first meta-analysis of physically active lessons. The effect on activity inside the lesson is one of the largest in the school-intervention literature, and it shrinks by most of an order of magnitude once you measure the whole day instead. Design governs the period itself, and its reach stops at the bell.

Activity during the lesson

d = 2.33 (1.42 to 3.25)

Activity across the whole day

d = 0.32 (0.18 to 0.46)

In-class educational outcomes

d = 0.81 (0.47 to 1.14)

Overall educational outcomes

d = 0.36 (0.09 to 0.63)

00.20.50.8

standard deviations

What it does not show: No effect was found on cognition or on health outcomes, but only three studies fed each of those pools, so the evidence there is simply thin. Most included studies were primary schools.

Norris, van Steen, Direito & Stamatakis (2020). British Journal of Sports Medicine, 54(14), 826–838. Read the record

04

Four in five adolescents worldwide do not move enough

298 surveys, 146 countries, 1.6 million students aged 11–17

The largest pooled measurement of adolescent activity ever assembled found 81.0% falling short of the hour a day. The gap between boys and girls was not closing: 77.6% of boys against 84.7% of girls, and the boys improved between 2001 and 2016 while the girls did not.

Insufficiently active, all adolescents

81%

Boys

77.6%

Girls

84.7%

Seven points worse than the boys, and the only one of the two that did not improve between 2001 and 2016.

What it does not show: This is self-reported survey data, which tends to overstate activity, so the true figures are likely worse. National figures vary widely and the global average should not be read as any one country’s number.

Guthold, Stevens, Riley & Bull (2020). The Lancet Child & Adolescent Health, 4(1), 23–35. Read the record

05

Indian adults have gone from a fifth to a half in twenty-two years

507 surveys, 5.7 million participants, 197 countries and territories

The prevalence of insufficient physical activity among Indian adults rose from 22.3% in 2000 to 49.4% in 2022, against a world figure of 31.3%. India now ranks 12th of 195 countries. Within India the gap between women and men is fifteen points.

World, 31.3%

22.3%

49.4%

2000

2022

Insufficient physical activity among Indian adults, against the world figure for the same year.

Within India, 2022

India, women, 2022

57%

India, men, 2022

42%

Fifteen points between them.

What it does not show: This measures adults, not schoolchildren, and it cannot attribute the rise to anything in particular. Urbanisation, work and transport all move these numbers. It establishes that the trend is bad and that it is worse here than in most places. It does not establish that physical education is the cause or the cure. The paper carries a December 2024 correction, which restored a missing author group and changed no figures.

Strain, Flaxman, Guthold, Semenova, Cowan, Riley, Bull & Stevens, with the Country Data Author Group (2024). The Lancet Global Health, 12(8), e1232–e1243. Read the record

06

The target is an average across the week, not a daily pass or fail

Global guideline, ages 5–17

At least an average of 60 minutes a day of moderate-to-vigorous activity, mostly aerobic, across the week, plus vigorous activity and work that loads muscle and bone on at least three days. The 2020 revision deliberately moved from "60 minutes daily" to "an average of 60 minutes a day", because a school week does not distribute evenly and the evidence never required that it did.

60 min

30

90

45

75

40

20

120

Mon

Tue

Wed

Thu

Fri

Sat

Sun

Four of these seven days fall under the hour. The week still averages 60 minutes a day, so it meets the guideline as written. This illustrates how the rule works; the paper reports no daily figures.

What it does not show: A guideline describes a population. It says nothing about what any individual child needs, and benefit keeps accruing past 60 minutes.

Bull et al., for the World Health Organization (2020). British Journal of Sports Medicine, 54(24), 1451–1462. The WHO 2020 guidelines. Read the record

Taken together

WHAT THIS EVIDENCE DOES NOT ESTABLISH

01

Almost all of this research was done outside India. The physiology travels; class sizes, facilities, climate and the shape of a school day do not.

02

Effects on academic marks are real but small. Any claim that sport will transform examination results is going beyond the evidence.

03

The strongest, most consistent findings are about behaviour, engagement and activity inside the lesson, which is exactly where lesson design has reach.

04

None of it demonstrates that software improves outcomes. It supports the case for well-designed physical education. The tooling then has to earn its own keep.

We would rather you knew all of that before you decided about us than after. The case for physical education is strong enough without being overstated, and the case for our tooling has to be made on its own.

What we actually builtWhat we are aiming at