Research Report · RR-01 · May 2026
The State of the Evidence
A comprehensive review of the cognitive, biological, affective and developmental neuroscience for the design of schools. Every principle carries a stated evidence tier and the one line basis for it, and where the evidence thins the paper says so rather than smoothing it.
The question it asks
What does the cognitive, biological, affective and developmental neuroscience say about the conditions that make learning possible, and what does that imply for how schools are structurally designed?
The organising lens is structural design: the architecture of schooling rather than the conduct of teaching. The timetable, the assessment cycle, the grouping decisions, the behaviour systems, the physical environment and the relational structures that condition every interaction within a school are the object of study, not curriculum or pedagogy, which sit downstream of them.
These structures are largely invisible in any given moment and feel like fixed conditions. Each was in fact a choice, and what was chosen can be chosen differently. Because structural choices persist and compound over years, getting them wrong is costly for a long time, and getting them right is correspondingly durable.
What it is not
It is a synthesis rather than a prescription. It sets out what the evidence is, tier by tier, including where that evidence is contested, inconvenient or absent, and it stops short of telling a school what to do. The companion white paper takes the editorial position and the recommendations; this paper establishes the ground they stand on.
It also declines a set of claims that circulate widely and do not survive contact with the literature. Where a popular idea is not supported, the paper says so and explains what the evidence actually shows instead, rather than quietly leaving it out.
How the evidence is graded
Three tiers, modelled on the discipline that mind, brain and education science has brought to the same problem. The system exists so that a reader can tell them apart, and the paper’s central discipline is that it does not blur them.
Robust
Converging support across populations, methods and decades, including in school populations, with mechanisms reasonably understood. Disagreement concerns refinement rather than existence.
Strong but contested
Well supported in direction, and actively debated in magnitude, mechanism or applicability. Enough to pilot carefully; not enough to promise.
Emerging
Theoretically grounded and promising, resting on thin or indirect evidence. Named as such so that nobody builds a policy on it by accident.
Amber marks the boundary where a tier drops, and it is used for nothing else anywhere in this work. A reader should always be able to see, at a glance, the point at which the evidence stops carrying the weight being put on it.
What is in it
Ten chapters in seven parts: the frame, the preconditions of learning, the social architecture, developmental variation, the digital environment, assessment, and what the whole of it will bear.
01
What this paper does, and what it does not
The single question the paper asks, the lens it uses, and the claims it declines to make.
02
How this paper treats evidence
The three tiers, how a principle earns one, and why the paper refuses to blur them.
03
Why structural design
Why the architecture of schooling, rather than the conduct of teaching, is the level at which the cognitive sciences have the most under-translated implications.
04
Biological preconditions
Sleep and circadian timing, where the misalignment between the adolescent body and the secondary day is as close to settled as the field gets; and physical activity, nutrition and the sensory environment of the building.
05
Cognitive mechanisms
Attention as the gating function, working memory as a hard constraint that instruction and assessment routinely overrun, and consolidation and spaced retrieval as a robust account of how durable learning forms.
06
Affective and social substrates
Belonging, psychological safety and peer dynamics: the conditions that determine whether those mechanisms are recruited at all.
07
Developmental variation
What changes across the years of schooling, from executive-function assembly in the early years to the social-affective reorientation of adolescence, and what varies between children of the same age.
08
The digital environment
The personal networked device treated as a structural-design variable rather than a behaviour problem, in four layers of very different evidential strength.
09
Assessment architecture
What a system chooses to measure, and how it reports it, propagating downward into what is taught and what children conclude about themselves; with British Columbia as the worked case.
10
Conclusion: what the evidence will bear
Which structural claims are robust enough to design around now, and which are not, stated as plainly as the evidence allows.
Where it stops
Why the paper stops short of recommendations
Two conclusions run through the whole of it. The first is that a small number of structural claims are robust enough to design around now. The second is that a larger set are not, and that the distance between those two groups is exactly where most educational advice goes wrong.
A synthesis that ends in a list of instructions has to flatten that distinction to produce the list. This one does not, which makes it less immediately usable and considerably harder to argue with.
The recommendations live downstream instead, in five structural projects that each take one failure and work it through to something a school can act on: the shape of the day, manufactured disadvantage, the assessment engine, the device, and designing for the developing brain.
See the five projectsThe report is free. Ask and we will send it.
Published May 2026, with corrections logged since. If you would rather talk it through than read 138 pages, that is what the advisory is for.
Request the reportAdvisory enquiry