Briefing
Who behaviour points fall hardest on
Classroom points systems are applied evenly and land unevenly. This briefing sets out how that happens, what the national data shows, what the Upper Tribunal said about it in February 2026, and the five things a school could do about it this term.
Written for governors, SENCOs, senior leaders and anyone who has been handed a behaviour policy and asked whether it is fair. It is designed to be printed.
The sixty-second version
Five things, if you read nothing else
- For some children, the behaviour being scored is the disability. Sitting still, eye contact, remembering equipment, arriving on time and not reacting to noise are features of a condition, not lapses in conduct, and they are precisely what a points system counts.
- Points are the front of a pipeline. Points become detentions, detentions become a record, the record becomes the evidence for a suspension. The national figures show exactly who that pipeline carries.
- A points system doesn't create bias. It converts it. A split-second judgement becomes a number on a board, and a number does not look like an opinion, least of all to next year's teacher.
- The reasonable-adjustments duty applies to behaviour policies. Confirmed by the Upper Tribunal in February 2026, covering sanctions and exclusions, not only admissions. Many policies predate that.
- Almost no school has looked at its own data. The app already holds a term's worth of evidence about who collects points and who doesn't. Breaking it down by SEND status, ethnicity and free school meal eligibility costs a filter.
For some children, the thing being scored is the disability
Sitting still. Making eye contact. Remembering the right book. Arriving on time. Not reacting to a noise everybody else has filtered out. For an autistic or ADHD child these are not lapses in conduct: they are the exact features of the condition, and they are exactly what a points system counts.
So the child is not being scored on their behaviour. They are being marked down for the way their brain works, several times a day, usually in front of an audience. And because the output is a number, it arrives looking like a fact about the child rather than a fact about the instrument that produced it.
The same applies at the quieter end. A child with anxiety, or one still acquiring English, may go a fortnight without doing anything a system would notice. Where a school gives only positive points and considers the fairness question settled, that child is not neutral. They are bottom, with no incident to discuss and no adult who thinks anything has gone wrong.
A score doesn't need a minus sign to be a ranking. Zero is only neutral on its own.
Points are the front of a pipeline
A point is not the end of the road. Points accumulate into detentions, detentions accumulate into a record, and the record becomes the evidence base for a suspension. We know with some precision how that pipeline distributes, because the government publishes it every year.
Permanent exclusion rate, England, 2022/23, per 100 pupils
| Group | Rate |
|---|---|
| White Gypsy or Roma | 0.43 |
| Traveller of Irish heritage | 0.35 |
| Mixed White and Black Caribbean | 0.24 |
| Black Caribbean | 0.18 |
| White British | 0.13 |
| All pupils | 0.11 |
| Black African | 0.06 |
| Indian · Chinese | 0.01 |
Source: DfE, Ethnicity facts and figures, permanent exclusions, 2022/23. The Gypsy, Roma and Traveller figures rest on small numbers and the department flags them as less reliable.
Note the shape of that table, because it is not the shape people expect. It does not say that minority ethnic pupils are excluded more. Some groups sit well below the national average and some sit several times above it, and the groups at the top have been at the top for as long as anyone has counted. Any argument that flattens this into a single claim will be taken apart by the first person who checks, which is why it is set out here in full.
The same pattern runs through need and through poverty, and there the gaps are wider still. In the Department for Education's termly figures for spring 2024/25, the suspension rate for pupils on SEN support was 9.79 per 100, against 2.35 for pupils with no identified special educational need. Children eligible for free school meals are more than four times as likely to be suspended, and more than five times as likely to be permanently excluded, as their better-off classmates.
None of this is caused by a points app. Exclusion disparities long predate any of these products. But points are where the record starts, they are the earliest visible expression of the same judgements, and they are the only part of the pipeline a parent can see before it becomes something more serious.
It doesn't create the bias. It launders it.
Two children do the same thing. One is read as having a difficult morning; the other as having an attitude. That judgement is made in about a second, by an adult with thirty other things happening, and the evidence is that it is not made identically about every child.
Dr Jahnine Davis's work on adultification bias describes what happens when notions of innocence and vulnerability are not extended to certain children, Black and Black mixed-heritage children in particular. Davis characterises it as a persistent act of dehumanisation, visible in the language used about children ("streetwise", "mature", "resilient") and in the substitution of suspicion for concern.
A points system does not invent any of that. What it does is convert it. A judgement call becomes a figure on a board; a figure does not look like an opinion. By half term it is a total, by July it is a record, and by the following September it is what the next teacher knows about that child before they have met.
The bias was always there. The app is what turns it into data, and data is much harder to argue with than a hunch.
The law has moved further than most behaviour policies
In UW v Cheam Academies Network [2026] UKUT 53 (AAC), decided on 4 February 2026, the Upper Tribunal held that a school's duty to make reasonable adjustments under the Equality Act 2010 extends to the operation and application of its behaviour policy, including sanctions and exclusions, not only to admissions. The case concerned a child who went through a run of detentions, then fixed-term exclusions, and then permanent exclusion.
The practical consequence is narrow but significant. Where a pupil is disabled (and autism, ADHD and a great many other conditions meet the Act's definition), the question is not simply whether the policy was applied consistently. It is whether an adjustment was actually considered at each stage, and whether the school can show that it was. Asserting that adjustments were thought about is not sufficient.
The Act does not require exempting disabled pupils from the rules, and nobody sensible is asking for that. It requires adjustments that are reasonable and proportionate. But a system that awards and deducts points automatically, uniformly and at speed is, by design, a system that does not pause to consider anything.
Treating every child identically is not the same as treating every child fairly. It only looks the same from the front of the room.
Five things a school could do this term
In rough order of effort. The first one is the one that matters, and it is also the cheapest.
Run the filter
Break one term's points data down by SEND status, ethnicity and free school meal eligibility. The app already holds it; this is an export and a pivot table. If the same groups sit at the bottom of every class list, the school now knows that in October rather than at an exclusion panel in June. If they don't, that is a genuinely good thing to be able to say.
An afternoonMake the totals private
Comparison is what gives a scoreboard its force, and it is also what does the damage. Scores visible to the child, their parent and their teacher keep almost all of the intended benefit. Scores on the whiteboard cost a great deal, and cost it disproportionately to the children who are always near the bottom of them.
A settings changeStop removing points already earned
Deducting an earned point says, retrospectively, that the good thing did not happen. It teaches nothing, it cannot be appealed, and it falls hardest on children whose difficulty is regulation rather than intent, who can lose a morning's work in one moment they could not control.
A policy lineWrite the adjustments down
If a child's plan says they need movement breaks, the behaviour policy has to say what happens when they take one. An adjustment held in a single teacher's head is not an adjustment; it is a kindness that expires in July when that teacher moves on. After February 2026, it is also the thing a tribunal will ask to see.
A termPut it in front of governors once a year, with the data attached
Behaviour systems are among the few things running on every child, every day, that are almost never formally reviewed. An annual item (what the system is, what evidence was looked at before adopting it, who it is falling on, and what changed as a result) is a modest ask and would put most schools ahead of the sector. On the evidence point, be ready for the answer to be that nobody reviewed any: the literature on these apps is small, mostly short-term, and weighted towards parent engagement rather than durable behaviour change, and the most-cited peer-reviewed paper on ClassDojo by name is a critique rather than a trial.
One agenda itemWhat this briefing is not
It is not an allegation against any particular school, and it does not claim that any school using a points app is discriminating. The national disparities described here have many causes and long predate this software.
It is also not an argument that teachers are the problem. Some are required to use these systems by a policy they had no part in writing. Some were never trained in anything else. Some have never worked in a school that didn't score children. And some find them useful, which is worth saying plainly rather than pretending otherwise: the case here is that working and being right are different tests.
The claim is narrower than that, and harder to dismiss: a system that scores children continuously, publicly and at speed will reproduce whatever judgements it is fed, and it will make them look objective. Any school running one is already holding the evidence of whether that is happening. Very few have looked.
Everything above, in full
- DfE / Ethnicity facts and figures, Permanent exclusions, 2022/23: exclusion rates by ethnic group.
ethnicity-facts-figures.service.gov.uk/education-skills-and-training/absence-and-exclusions/permanent-exclusions/latest - nasen, Rising suspensions and steady exclusion rates: what the latest data means for schools: suspension rates by SEN status, spring term 2024/25.
nasen.org.uk/news/rising-suspensions-and-steady-exclusion-rates-what-latest-data-means-schools - School Exclusions Hub, Inequality in school exclusions: who's losing out?: free school meal and ethnicity ratios, 2023/24.
schoolexclusionshub.org.uk/inequality-in-school-exclusions-whos-losing-out - UW v Cheam Academies Network (SEN) [2026] UKUT 53 (AAC), 4 February 2026: reasonable adjustments and behaviour policies.
gov.uk/administrative-appeals-tribunal-decisions/uw-v-cheam-academies-network-sen-2026-ukut-53-aac - Jahnine Davis, Adultification bias within child protection and safeguarding, HM Inspectorate of Probation, Academic Insights 2022/06.
- Manolev, J., Sullivan, A., & Millei, Z. (2019). The datafication of discipline: ClassDojo, surveillance and a performative classroom culture. Learning, Media and Technology, 44(1), 36–51.
- Mittiga, S. et al. (2024). A systematic review of classroom behaviour management apps. Education and Treatment of Children, 47, 83–104: 15 peer-reviewed studies of apps in this category (ClassDojo included, treated as a token economy); only 53% met quality standards, and 3 of the 15 measured any learning outcome. The review's own conclusion: mixed and preliminary.
- Sherman, L. E., Payton, A. A., Hernandez, L. M., Greenfield, P. M., & Dapretto, M. (2016). The power of the like in adolescence. Psychological Science, 27(7), 1027–1035.
- Lindström, B., Bellander, M., Schultner, D. T., Chang, A., Tobler, P. N., & Amodio, D. M. (2021). A computational reward learning account of social media engagement. Nature Communications, 12, 1311.
- Education Endowment Foundation (2019), Improving Behaviour in Schools, guidance report: reward systems as part of a broader strategy; UK research on reward-system-only interventions inconclusive.
educationendowmentfoundation.org.uk/education-evidence/guidance-reports/behaviour - Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6), 627–668.
If you are the one asking
The question that starts it
"Can you show me the points data broken down by SEND status, ethnicity and free school meal eligibility?" It costs a school nothing to answer, and if the answer is that nobody has ever looked, that is itself the finding.
The take action page has the rest of the questions, and who to put them to.