What rising suspensions in schools tell leaders early

Suspensions in schools in England have risen sharply since the pandemic, and by the time a rise shows in your figures, the pattern behind it has usually been building for terms. Trust leaders and headteachers therefore face a choice. They can keep relying on suspension counts, which are accurate but arrive after the event, or they can add a measure that shows where pressure is building while there is still time to change the outcome.

This article sets out what drives suspensions and why behaviour data alone reaches leaders late. It then looks at how belonging data, read by cohort rather than by whole school, can point to the groups that need attention first, and how a low score can become an action with a named owner.

How suspension figures in England have moved and what they count

A suspension, formerly called a fixed-period exclusion, is a removal from school for a set period. A permanent exclusion removes a pupil from the school roll. The Department for Education (DfE) moved to the term "suspension" in its statutory guidance on suspension and permanent exclusion, which took effect in 2023, and the wording in this article follows it.

The national picture comes from the DfE's annual statistical release, Suspensions and permanent exclusions in England, which is published each summer and covers the previous academic year. Suspensions have risen in successive releases since the pandemic and now sit above pre-pandemic levels. Leaders who want the current headline rate should take it from the latest release, as the figure moves with each publication.

The same release shows consistent differences between groups. Suspension rates are higher for pupils with SEN support or an education, health and care plan, for pupils eligible for free school meals, and for some ethnic groups, including pupils of Gypsy/Roma and Traveller of Irish Heritage backgrounds. These gaps have appeared year after year, so they are better read as a structural pattern than as a one-off spike.

The data has a built-in limit. It records what happened, such as how many suspensions, for which reason and for which pupils. It says nothing about the experience that came before, and it reaches trustees months after the events, often as a termly summary or an annual release. A board can see that suspensions went up in one academy last year. It cannot see from that figure which pupils were already pulling away two terms earlier.

Why suspensions in schools are the end of a longer sequence

Many school leaders describe a recognisable run of events. A pupil starts to feel disconnected from the school. Lateness and patchy attendance follow, then low-level disruption, then repeated sanctions, and finally a suspension. This is a commonly described pattern, not a proven sequence that every pupil follows, and some suspensions follow a single serious incident with no build-up. Still, it matches what many pastoral teams see when they look back through a pupil's record.

It helps to separate proximate triggers from underlying conditions. The trigger is the incident or the sanction threshold that was crossed on a particular day. The underlying conditions are quieter: whether the pupil has a good relationship with at least one adult, whether they feel safe in corridors and at lunch, and whether they feel the school is a place for people like them. Behaviour systems deal well with triggers. They are less well placed to see conditions, because conditions do not generate a log entry.

That gap explains a common misconception, which is that a low suspension rate means a school is well run. A low rate can sit alongside high persistent absence, which is the term for a pupil missing 10% or more of possible sessions. It can also sit alongside off-rolling risk, or alongside a group of pupils who have simply stopped engaging without causing a problem. A school with a high rate may be dealing openly with issues that another school has not yet surfaced. Reading suspensions in isolation can flatter one academy and unfairly mark down another.

For that reason, many trusts read suspensions alongside attendance and a measure of how pupils experience school. The case for including belonging has support in the research literature, though any single study should be checked against its own methods and year before it goes into a board paper. The practical point does not depend on one finding. If the conditions come first and the behaviour follows, the earlier measure is the one that describes the conditions.

What belonging data can show before behaviour data does

Pupils' answers about feeling safe, valued, listened to and connected can shift before attendance and behaviour records do. A pupil who has begun to disengage may still be arriving on time and following the rules. Their responses to a short set of structured questions can show a change that no register or behaviour log has yet picked up. Any claim about exactly how much earlier needs a named source behind it, and leaders are right to ask for one.

The design of the questions matters as much as the idea. A belonging survey usually covers a handful of topics, such as relationships with staff, feeling safe, peer connection, fairness, a sense of being represented, and engagement with learning. Each is measured with structured items. A typical pupil question is a rating scale against a statement such as "Adults in this school listen to me", with options running from strongly disagree to strongly agree. Pupil surveys of this kind work best without free-text boxes, which keeps responses comparable and avoids pupils disclosing sensitive information in a place nobody is monitoring in real time.

A few practical choices tend to separate surveys that get used from those that do not:

Length is the first. A survey of 5 to 10 minutes can be completed in a tutor period, which helps participation, particularly among the pupils leaders most want to hear from. The second is consistency. When every school in a trust asks the same questions at the same point in the term, the results can be compared. Anonymous surveys tend to suit a first baseline, where honesty matters most, while named surveys let a school follow up with individual pupils or build cohort views from MIS data. Trusts choose according to purpose, and tell pupils which it is.

Access needs planning too. Pupils with EAL may need the questions in their home language or read aloud, and pupils with SEND may need adapted formats or adult support. Without that, the results reflect the pupils who found the survey easiest, which is rarely the group at greatest risk.

Response rate is worth reading as a finding in its own right. If participation in one cohort is noticeably lower than elsewhere, that gap deserves a follow-up conversation, because it may reflect access problems or disengagement. A benchmark for a good response rate should come from a named source, not from habit.

Reading the results by cohort rather than by whole school

A whole-school average describes the typical pupil, and the pupils at greatest risk are often not typical. A worked example shows how this plays out. The figures below are hypothetical and are for illustration only.

Suppose a trust has five secondary academies with broadly similar suspension rates. Academy C reports a healthy overall belonging score, close to the trust average. When the trust breaks the results down, Year 8 pupils eligible for Pupil Premium in Academy C score well below the trust average on the statement "I feel I belong at this school" and on items about fairness. Their peers in other year groups score comfortably. The headline score hid the gap because the group is a small share of the school, and the rest of the pupils pulled the average up.

Nothing in Academy C's suspension data would have flagged this yet. The cohort may be sitting a term or two before the sequence described earlier reaches its later stages. Breaking results down by year group, SEND, Pupil Premium, EAL, free school meals and combinations of these is how such a group becomes visible. Intersections matter in particular, because a pupil with SEND who is also eligible for free school meals may have a very different experience from either group taken alone.

Comparison between academies only works when the method is identical. The same questions, asked in the same way and at the same time, mean the board sees what pupils, staff and families reported, not a headteacher's summary of it. Summaries are useful context, but they pass through a layer of interpretation that makes like-for-like comparison hard.

Small cohorts need care. A group of six pupils can swing sharply on one or two responses, and reporting at that level can identify individuals. Many trusts set a minimum group size below which results are suppressed or treated cautiously, which also protects anonymity under UK GDPR. The right threshold is a local decision, and it helps to agree it before the first results arrive. A low score in a small group is better treated as a prompt for conversation than as a conclusion.

Turning a low-scoring cohort into an intervention with an owner

A finding changes nothing until someone is responsible for it. Many trusts work through the same termly cycle, which has five stages:

1. Measure: run the same short survey for pupils, staff and families.
2. Segment: break results down by cohort to find where scores diverge from the average.
3. Connect: set belonging results beside attendance, behaviour and MIS data for the same groups.
4. Respond: turn each low-scoring cohort into a named intervention group with an owner.
5. Assure: report the cohort, action, owner and review date to leaders and trustees, then measure again.

Return to Academy C. Its pastoral lead and a named senior leader meet the Year 8 tutors, share the pattern in the scores, and talk through what tutors are seeing day to day. They agree two or three changes. One might be mentoring links between the Pupil Premium group and specific adults. Another might be a review of how sanctions are applied for this cohort, looking at whether the same behaviour draws different responses across classrooms. They set the review point at next term's survey, so the group's scores have a defined moment to be re-read.

For trustees, the useful report is short. It names the cohort, the action, the owner and the next measurement date. Framed this way it reads as planning, a record of where the trust chose to put attention, and not as a defensive account prepared for an inspection.

Belonging data does not replace the rest of the picture. It sits alongside MIS, attendance and behaviour data and the trust's existing behaviour strategy, and changes are best judged against all of them across several terms. One survey result is a signal. A direction of travel over three or four terms, supported by attendance and behaviour trends, is evidence.

How Satchel Pulse can help

Perceptions, from Satchel Pulse, runs termly belonging surveys for pupils, staff and families. Cohorts such as year group, SEND, Pupil Premium and EAL are generated from the MIS via Wonde, and a trust-wide dashboard puts every academy on one comparable view. Satchel Pulse's own data suggests belonging results can flag cohorts at risk of rising absence and suspensions up to 18 months before those patterns show in attendance or behaviour figures. When a cohort scores low, the library of 6,500+ intervention resources gives leaders somewhere to start, and the results can support the evidence kept on inclusion and belonging.

Suspension data tells you where pressure has already surfaced. Belonging data shows where it is building. The step for this term is to schedule a baseline survey before the end of term, choose one cohort to look at first, name an owner, and put the review date in the spring calendar.

Learn more about our services