Skip to content
The Briefing: five minutes on what moved, every Monday. Subscribe free Editorial standardsCorrectionsContact
Questions  /  06 · Development

Question 06 · A living dossier

How should school change as a child grows?

Updated 25 September 2026Version 1.0Stands on The State of the Evidence, chapters 4, 6 and 7By Shilpa Kapur SinghCite this

The current answer

School is built in phases, and children fall through the joins. Redesign the four seams, at 4, 11, 16 and 18, starting with the move into secondary, where the demands jump just as the adolescent brain is recalibrating. And refuse the story of the closing window. Most windows narrow. Very few close.

A child who has not learned something by a given age can still learn it. Designing as though they cannot turns a slower start into a fixed outcome.

Four seams across a changing brain

Four seams across a changing brainSchool phase changes at ages 4, 11, 16 and 18, set against executive function assembling from about 3 to 7, working memory growing fastest from about 10 to 15, peak onset of mental health conditions at 14.5, and adolescence running to about 24.510152025Age4 · Starts school11 · Moves to secondary16 · Sits exams18 · LeavesExecutive function assembles, about 3 to 7Working memory grows fastest, about 10 to 15Adolescence, by many definitions, to about 24Peak onset of mental ill-health, 14.5
Ages are approximate and describe averages, not individual children. School ages are for England.
  1. 4 · Starting schoolSchool begins while executive function is still assembling, roughly between three and seven. The youngest in the year are the furthest behind in it.
  2. 11 · The move to secondaryMore teachers, more rooms and more to hold in mind, just as working memory is in its fastest growth. Too often the move also brings less autonomy and fewer close adults, when young adolescents need more. Eccles et al., 1993
  3. 16 · Terminal examsThe highest-stakes assessment lands close to the peak age of onset for mental health conditions. Solmi et al., 2022: 192 studies, over 700,000 people
  4. 18 · Participation endsCompulsory education stops while adolescence, by many definitions, runs on to about 24.

Variation is the design brief

The research project
718,838

education, health and care plans in England in January 2026, up from 576,474 two years earlier.

DfE, 2024 to 2026
~2/3

of the variation in which pupils are identified with special educational needs is linked to the school they attend, not to the pupils.

Education Policy Institute
5.9%

of children meet the criteria for ADHD. Variation on that scale is a design parameter, not an exception to plan around.

International consensus statement on ADHD

What to do first

01
Audit the seams
For each of the four joins, write down what changes on the first day: the adults, the demands, and what information travels with the child. Most schools have never put the four side by side.
02
Give every child a named adult across the move to secondary
One person who knows them before the move and stays with them after it. It is the cheapest protection at the most exposed seam.
03
Name the layer
When a child struggles, ask which layer explains it: their stage of development, the school’s design, or a label. Each needs a different response.
04
Design around dimensions, not labels
Plan for the range of attention, working memory and self-regulation in every class, rather than waiting for a diagnosis to trigger support.

What would change our mind

  • Evidence that children’s needs are met as well across the joins as within a phase.
  • A sensitive period relevant to schooling that genuinely closes, rather than narrows.
  • Evidence that diagnostic labels predict what support a child needs better than their measured profile does.

Where this comes from

What changed

  • 25 Sept 2026
    Published as a living dossier, version 1.0.
  • Sept 2026
    Evidence brief EB-005 and white paper WP-005 published, after three corrections to the report.
  • May 2026
    First published in The State of the Evidence.

Also in the evidence

On second-language grammar, the largest dataset, 1.1 million people, can be read two ways: a sharp decline at the end of adolescence, or a continuous one. The question between them is open. The claim school design is usually asked to act on, a high plateau and then a cliff, is the one the evidence will not carry. Chen & Hartshorne, 2021; Slik et al., 2022

How to cite

Kapur Singh, S. (2026). How should school change as a child grows? Living dossier, version 1.0, 25 September 2026. The EdJournal.