Where Does School Avoidance Actually Sit?
Updated: Sep 23
Understanding school avoidance through the interaction between the child, their development and the environment.
School avoidance: an outcome, and then part of what happens next
A register tells us whether a child was recorded as attending school. On this morning, this child was not. What it does not tell us is why.
That distinction sounds obvious and is repeatedly lost. School avoidance describes an outcome. It does not establish anxiety, demand avoidance, sensory overload, low motivation, bullying, trauma or unmet learning need. Each may become a reasonable hypothesis. None is established by the fact of absence.
The scale is not in dispute. Across 2024/25 in England, 18.14% of pupils were persistently absent — down from 19.95% the year before, but well above the pre-pandemic rate of 10.86%. Severe absence moved the other way, rising to 2.39% from 2.30%, against 0.85% in 2018/19. Pupils with special educational needs were almost four times more likely to be severely absent than those with no identified need, and that gap widened.
There is a second point, which matters as much. Absence does not stay an outcome. Once a pattern establishes, it changes the conditions around the child. Learning is missed. Peer relationships may change. Work can accumulate. Returning becomes more conspicuous. Formal processes engage. Family life often rearranges itself around getting through the morning. Current statutory guidance makes a version of this point directly, observing that a prolonged period of absence may heighten anxious feelings about attending in future. The causal arrow does not run one way only.
Attendance matters, and nothing here disputes that. The question is what follows from that: if attendance matters, what gives this child the best chance of achieving and sustaining it?
What do we actually know about this child's experience?
Sit in enough meetings and the same sentences recur. He's fine once he's in. She's doing it for attention. It's the anxiety. It's her PDA. They know if they make enough fuss they'll get to stay home.
Any of these might contain something useful. None is established by the behaviour that prompted it. Each is an inference about an internal state, made by an adult, from what they could see. Inferences are necessary — we cannot work without them — but they are hypotheses to be tested, not findings to be adopted.
This is not pedantry about language. A systematic review of school staff's views on non-attendance found that staff conceptualise the problem in markedly different ways, including within-child, home-related and school-related factors, and that these conceptualisations may influence how staff interpret and respond to non-attendance. Where we locate the problem determines what we try to change. Understand it as anxiety and we target anxiety. Understand it as motivation and we reach for incentives. Understand it as sensory and we alter the environment. Understand it as developmental mismatch and we change the demand and the scaffolding. None of these should be selected because a label has been applied.
The child's own account is a different kind of evidence, and we treat it too casually. A professional saying she is anxious about separation is inferring. A child saying everyone looks at me when I go back is reporting. That account has limits — a child may describe the most recent bad day rather than the pattern, may give the answer they think is wanted, may not have the words — but those are the limits of testimony, not grounds for replacing it with our own version. Families are in the same position. A parent who sees how long recovery takes on a Tuesday evening holds information nobody in the building can observe.
And sometimes a child says: I don't know why. I just can't go in. The absence of an explanation from the child is not permission for adults to manufacture one. It tells us we do not yet understand enough.
One further caution. Emotionally based school avoidance (EBSA) can be useful descriptive language, but the label does not establish what is driving the difficulty for an individual child — and not all non-attendance is EBSA in any case. Illness, caring responsibilities, bullying, transport and cost, safeguarding concerns and a placement that was never right all produce the same register entry.
What are we actually asking this child to do?
"Go into school" sounds like a single behaviour. It is not.
Depending on the child and the day, it may mean waking and organising themselves, tolerating a journey, separating from a parent, entering a building that is loud before anything has begun, remembering equipment, holding a multi-step instruction, switching attention repeatedly, starting work without being certain what is wanted, asking for help in front of others, tolerating a mistake, managing twenty unstructured minutes at lunch with no assigned role, and doing it again tomorrow.
Those requirements are not interchangeable, and this is where generic thinking fails. A child may manage the noise of the dining hall perfectly well and find the unpredictability of entering it intolerable. Another may cope with the academic demands of lessons and be exhausted by the social navigation between them.
For autistic children, sensory processing, communication, uncertainty, social interpretation and the executive demands of the school day may all be relevant. But autism does not tell us which of these matters for this child, in this situation. Diagnosis should inform the questions we ask, not answer them in advance.
The useful question is therefore not whether a child of this age should manage all this. It is: which parts of it can this particular child currently do, under which conditions, with what support, and at what cost?
That reframing needs handling carefully, because it is easy to turn it into a new way of locating deficit. Breaking a demand down does not establish that a child has an executive function deficit or a working memory problem. It tells us what the task appears to require, and gives us hypotheses to investigate. The distinction matters technically as well as ethically: performance-based measures and rating measures of executive function correlate poorly with each other, with a median correlation of around .19 across the studies reviewed, which suggests they are not straightforwardly indexing one underlying thing. A cognitive construct inferred from behaviour is still inferred from behaviour. Replacing he's avoiding demands with he has working memory difficulties upgrades the vocabulary, not necessarily the evidence.
Readiness for a particular demand therefore cannot be read from the child alone. It depends on the capabilities the child brings, the nature of the task, and the conditions and support under which the task is presented. A child may be entirely unready to write a paragraph independently and entirely ready to write one with a model, a scribe for the first sentence and a clear finishing point. "Not ready" specifies what is required; it is not a verdict about a person.
This is not an argument for lower expectations. Readiness thinking has been criticised, with reason, for placing the problem inside the child and justifying doing less, and the safeguard matters: adjusting a demand must not mean permanently removing the opportunities through which the capability would develop.
When demands and available capability don't match
What happens when what is asked exceeds what is available, repeatedly, across a school day?
It may not appear first as absence. It may appear as effort. A child manages, but at a cost invisible at school and obvious at home: longer to settle afterwards, shorter fuse, less appetite, Sunday evenings that deteriorate. Then participation narrows — present, but no longer putting a hand up, no longer going to the club, arriving late to the hardest lesson. Then avoidance of specific parts of the day. Then acute distress at the door. Then non-attendance. Then, sometimes, a fragile partial return treated as resolution.
That sequence needs care. It is a pattern, not a law, and certainly not a scale to score a child against. Many children show effort, withdrawal and long recovery and go on attending for years. Read backwards, after attendance has collapsed, almost anything looks like a warning sign.
A second hazard sits alongside that one, and the research illustrates it well. A study using England's statutory early years assessment, drawing on pre-pandemic data for more than sixty thousand children in one English district, found that children assessed as not "school ready" at entry had higher odds of later persistent absence, even after accounting for special educational needs, socioeconomic circumstances, sex, ethnicity and English as an additional language. The association matters, and it supports the possibility that early mismatch has downstream consequences. But note the measure: a child rated as emerging on any single early learning goal counts as not ready. That binary predicts risk across a population while telling us almost nothing about which demands were hard for which child, or what would need to change for a nine-year-old now. A risk marker identifies a group worth attending to. It does not explain an individual.
Most importantly, mismatch does not automatically mean the demand is wrong. Sometimes the scaffolding is missing and needs building. Sometimes a reasonable, developmentally appropriate skill has never been taught, and teaching it is exactly right. Sometimes the environment generates difficulty no amount of individual skill should be required to absorb. Sometimes the expectation itself is inappropriate. Often several are true at once. Deciding in advance which it will be is the mistake.
What capabilities does the environment need to respond?
We routinely ask what skills a child needs in order to cope with school. The question can be turned around.
Work on capable environments has asked us to examine the social, physical and organisational characteristics of the support someone receives, rather than concentrating on the person whose behaviour concerns us. That work was developed in community services for people with learning disabilities, not in schools; bringing the question into education is my application of it, not something the original literature established. But I think the question transfers well: what capabilities does this environment need in order to support this child?
Some answers are familiar — communication, predictability, relational consistency, sensory conditions, genuine choice, recovery opportunities, consistency between adults. One organises the rest.
The capability to notice is not the same as the capability to respond.
Statutory guidance already requires schools to monitor attendance closely, identify early indicators, and look beyond persistent and severe absence at all levels of absence so pupils can be supported before patterns entrench. The detection duty exists. Arguing for earlier identification is arguing for something already mandated. The harder question is what happens next. A system can detect deterioration accurately and still fail a child if nobody has the time or expertise to work out what the pattern means, and no route exists to change anything as a result. Noticing the problem is of limited value if the system does not have the capability to respond to what it has noticed.
That capacity is unevenly distributed, and it would be dishonest to pretend otherwise. Some of the list above sits within a school's discretion and could change next week. Some is constrained by funding, staffing and a statutory assessment system under visible strain — 718,800 children and young people held education, health and care plans in England in January 2026, a 12.5% rise in a year and the largest annual rise since plans began. Pretending the second category does not exist is its own form of blame.
Is the support provided the support the child experiences?
A school can often demonstrate a great deal: a plan, adjustments, a safe space, a named adult, a visual timetable, reduced hours, sensory equipment. Each may be appropriate and well intentioned.
The existence of provision is not evidence of the experience or effectiveness of that provision.
A named adult who teaches four other classes is available in principle. A visual timetable can increase predictability while still displaying a transition the child cannot currently manage. Reduced hours can remove time without removing the part of the day that was driving the difficulty. Work on the acceptability of healthcare interventions is useful here — and the transfer into education is mine — because it asks us to distinguish between support that looks acceptable in advance and support as the person actually experiences it. It also asks us to consider the burden of accessing that support and what it costs the person to use it.
None of this implies that schools are being disingenuous. The gap is structural, not deceptive. It opens because provision is recorded at the point of decision and rarely examined at the point of use.
Which suggests what to ask. Not whether the support exists, but what happens when the child needs it: who they have to approach, what they have to say, how long it takes, and what it costs them. That means asking the child and treating the answer as data rather than as a view to be noted. This is not an optional refinement. The Children and Families Act requires regard to the views, wishes and feelings of the child and their parents, and to the importance of their participating as fully as possible in decisions, with the information and support that makes participation possible. Attendance guidance reinforces it, expecting adjustments to be agreed with, and regularly reviewed with, the pupil and their parents.
From understanding to action
None of this yields a programme to apply. A UK systematic review of school-partnered approaches to emotionally based school avoidance found too few methodologically adequate studies to determine whether interventions work; a wider systematic review of interventions for persistent absence found variable effects, with favourable outcomes reported for some mentoring, family-involvement and school-counselling approaches, but an evidence base that remains limited. The absence of an established package does not mean nothing works. It means formulation carries weight a protocol cannot. So what does that actually look like?
Start smaller than the problem. Do not try to formulate "school". Choose one specific, difficult part of the day: the gate at 8.40, Monday maths, the transition into lunch, the first morning back after an absence, the changing rooms. Then work through five questions. The sequence draws on familiar functional assessment principles, with one addition that is too easily skipped: what, exactly, are we asking this child to do?
What actually happens? Describe it without explaining it. When, where, with whom, how often. What does the child say? What do family and school each observe?
What is the child being asked to do? Break the situation down. What does successful participation actually require, step by step?
Which parts are currently difficult, and under what conditions? Not what deficits the child has. What differs between good days and difficult ones? What can they do with support that they cannot yet do without it?
What happens around the difficulty? What precedes it, what follows it, what is the child moving towards or away from, and how long does recovery take? Establishing that avoidance is escape-maintained — that getting away from the situation is what keeps the avoidance going — identifies a contingency; it does not tell us why the situation became aversive enough that escape was worth having. That is where the work starts, not where it stops.
What can we change, and what happens when we do? Skill, demand, communication, teaching, relationship, timing, sensory conditions, predictability, access to help. Agree one or two changes, say in advance what you expect them to affect, and agree what you will look for afterwards. Stating the expectation first is what makes this iterative rather than merely descriptive: if the reasoning is wrong, you want something that will tell you so.
Here is the shape of it. Take the safe space that looks good on paper. A child explains that they cannot ask for it when everyone is looking. So we ask what accessing it currently requires: noticing their own escalation, finding the adult, approaching them publicly, producing speech while distressed, explaining themselves, waiting for permission, then leaving in front of the class. "Access to a safe space" turns out to contain seven demands. Observation may then show that access works when a familiar adult notices early and offers it quietly. That gives a testable change: can access be made available through an agreed signal, without requiring a public verbal request?
Then we try it. A change is a test, not a solution, and it should be treated as one — including changes that teach the child something new, which are sometimes exactly right and sometimes a way of asking a child to tolerate conditions that should not be tolerated.
How will we know? Attendance is important data. It is not, on its own, evidence that support is working. A child can be in the building and overwhelmed, in the lesson and unable to learn, compliant and profoundly distressed, present all week and unavailable to their family all weekend. So we watch attendance alongside learning, participation, relationships, autonomy, distress and how long recovery takes — domains drawn from quality-of-life work designed to be assessed through the person's own account as well as others' reports — and we ask the child. Where a temporary reduction in hours is part of the plan, current guidance sets the terms clearly: exceptional, agreed with the parent the pupil normally lives with, part of a wider support, health care or reintegration plan, with regular reviews involving the pupil and a proposed end point. A planned pathway, not an unexamined destination.
Where the scope for change sits
If school avoidance is understood as a problem located inside a child, the number of things we can change turns out to be very small. Once we examine the interaction between what a child is being asked to do, what they can currently do, and the conditions under which they are asked, considerably more becomes changeable.
The child has agency, a perspective worth more than we usually grant it, and a right to participate in decisions about their own education. Adults hold disproportionate control over the timetables, demands, relationships, environments and expectations surrounding them, and therefore disproportionate responsibility for changing those conditions.
The question is no longer simply how we get this child back into school. It is what we need to understand, teach, support or change to make meaningful and sustainable participation possible. Understanding the interaction does not reduce what we can do. It expands it.
David Corcoran is a Chartered Psychologist and Director of Beyond Thrive, working across autism, learning disability, complex support needs and the systems that surround them.
References and notes
Department for Education, Pupil absence in schools in England: Academic year 2024/25, published 26 March 2026, updated 21 May 2026. Accredited official statistics.
Department for Education, Working together to improve school attendance, July 2026, paragraph 54. Statutory guidance applying from September 2026.
Alaimo, G. and Kelly, C. (2025), ‘School staffs’ views on student non-attendance: a systematic literature review’, Frontiers in Education, 10, 1599065. https://doi.org/10.3389/feduc.2025.1599065. Synthesis of eleven qualitative and mixed-method studies.
Toplak, M. E., West, R. F. and Stanovich, K. E. (2013), ‘Practitioner review: do performance-based measures and ratings of executive function assess the same construct?’, Journal of Child Psychology and Psychiatry, 54(2), 131–143. https://doi.org/10.1111/jcpp.12001. Median correlation of .19 across 20 studies.
Carlton, M. P. and Winsler, A. (1999), ‘School readiness: the need for a paradigm shift’, School Psychology Review, 28(3), 338–352. https://doi.org/10.1080/02796015.1999.12085969. US early-years literature.
Wood, M. L., Gunning, L. and Mon-Williams, M. (2024), ‘The relationship between “school readiness” and later persistent absenteeism’, Royal Society Open Science, 11(6), 240272, published 26 June 2024. https://doi.org/10.1098/rsos.240272. Study of 62,598 children using the Connected Bradford linked dataset, academic years 2012/13–2019/20.
McGill, P., Bradshaw, J., Smyth, G., Hurman, M. and Roy, A. (2020), ‘Capable environments’, Tizard Learning Disability Review, 25(3), 109–116. https://doi.org/10.1108/TLDR-05-2020-0007. Conceptual elaboration, not an efficacy study.
Department for Education, Working together to improve school attendance, July 2026, paragraph 44.
Department for Education, Education, health and care plans: England 2026, reporting year 2026, published 25 June 2026, updated 30 June 2026. Accredited official statistics.
Sekhon, M., Cartwright, M. and Francis, J. (2017), ‘Acceptability of healthcare interventions: an overview of reviews and development of a theoretical framework’, BMC Health Services Research, 17(1), 88. https://doi.org/10.1186/s12913-017-2031-8.
Children and Families Act 2014, section 19. See also Department for Education and Department of Health, Special educational needs and disability code of practice: 0 to 25 years, January 2015, paragraph 1.1.
Department for Education, Working together to improve school attendance, July 2026, paragraphs 58–59.
McDonald, C. and Pervez, A. (2025), ‘School partnered approaches to emotionally based school avoidance in UK primary and secondary school-age learners: a systematic review’, British Educational Research Journal, 51(6), 2871–2892. https://doi.org/10.1002/berj.4205.
Middleton, A., Watson, M. and Anderson, J. K. (2026), ‘What school-based interventions work to improve attendance in secondary school students with persistent absence? A systematic review’, Frontiers in Child and Adolescent Psychiatry, 4, 1603680, published 12 January 2026. https://doi.org/10.3389/frcha.2025.1603680. Sixteen studies, mostly from the United States with one Australian study; study quality ranged from weak to moderate.
Schalock, R. L. and Verdugo, M. A. (2002), Handbook on Quality of Life for Human Service Practitioners. Washington, DC: American Association on Mental Retardation. Eight core domains, assessed through self-report and the reports of others.
Department for Education, Working together to improve school attendance, July 2026, paragraphs 66–67.
Discuss your situation
If school avoidance is becoming difficult to understand, or existing support is not producing change, an initial conversation can help clarify what may need closer attention. You can book a free 30-minute conversation to discuss the situation and whether Beyond Thrive is likely to be relevant.
Comments