Pupil engagement: what it is and how to measure it

Attendance can be steady and behaviour logs calm while a Year 8 class sits through lessons with nothing going on behind the eyes. Pupil engagement is a different thing from presence and compliance, and the two standard datasets only catch it once it has turned into absence or incident. Leaders who want to see it sooner need to ask pupils directly, in the same way, every term.

This article sets out what pupil engagement means in practice, what attendance and behaviour data miss, which questions give usable answers, and how a trust can compare results across academies and act on them.

Three kinds of engagement, and why they get confused

The most widely used framework comes from Fredricks, Blumenfeld and Paris, published in the Review of Educational Research in 2004. It splits engagement into three parts. Behavioural engagement is participation: attending, following routines, taking part in lessons and school life. Emotional engagement is how a pupil feels about school, teachers and peers, including interest and a sense of connection. Cognitive engagement is the mental effort a pupil puts into learning, such as persisting with difficult work and wanting to understand rather than just finish.

The three do not move together. A pupil can behave well, feel indifferent and do the minimum. Another can be noisy, challenge the teacher and be deeply absorbed in the topic. This is why compliance is a poor proxy. A quiet, present, punctual pupil may be disengaged on two of the three dimensions, and nothing in the behaviour log will say so.

Engagement and belonging are not the same thing

Belonging is the sense of being accepted and valued by the people in a school. It sits underneath emotional engagement as a driver. A pupil who feels that adults know them and peers accept them has a reason to invest effort. A pupil who feels like an outsider may still turn up and comply, but the investment is thin. Measuring belonging and engagement separately is useful, because a low belonging score tells you where to look for the cause of flat engagement.

Engagement is a response, not a trait

The most costly misconception is that engagement is a fixed property of the pupil: some children are motivated and some are not. The framework treats it as responsive to context. The same pupil can be fully engaged in one teacher’s lessons and withdrawn in another’s. Relationships, task design, routines and school culture all move it. That is good news for leaders, because it means engagement can be changed, and it is the reason to measure it at cohort and class level rather than labelling individuals.

What attendance and behaviour data can and cannot show

Attendance, suspensions and behaviour points are lagging indicators. They record disengagement after it has become visible in what a pupil does. A leading indicator is something that tends to move earlier, which gives you time to respond. Lagging data remains essential for accountability and for spotting individual pupils in difficulty. It is a weak tool for seeing what is building.

Consider an illustrative pupil in Year 8. She attends at 96%, has never been sanctioned and hands work in on time. In a survey she disagrees with the statement that someone would notice if she were absent. Nothing in the MIS flags her. If a third of her year group answered the same way, the school would have a pattern that no register or behaviour report would show for months, if ever.

The wider picture explains why trusts care about catching this early. The Department for Education publishes pupil attendance and suspension statistics for England, and persistent absence has remained a headline concern in recent releases. Check the latest release on GOV.UK for current figures, since they update through the year and by term. For a trust, persistent absence is a board-level measure, which makes any earlier signal worth having.

Perception data adds the pupil’s own account. It can surface cohort-level patterns, such as one year group or one academy scoring lower on feeling part of the school, before they show in attendance or behaviour. Treat this as a hypothesis to test, not a guarantee. The sensible approach is to run surveys for a few terms, compare the results with your own attendance and behaviour data, and see how closely they line up in your schools. Some patterns will show up in both. Some will not, and that is also useful to know.

What to ask pupils, staff and families

Good engagement questions ask about specific experiences rather than general satisfaction. For pupils, that means items on whether they feel listened to, whether they have an adult in school they could go to, whether lessons feel worth their time, and whether they feel part of the school community. Pupils can answer these concretely. “Do you like school?” is too blunt to act on.

Keep the survey short. Five to ten minutes is realistic in a tutor period or lesson, and short surveys get higher completion. Use the same questions each term so you can compare like with like, and adjust the language for the age group. Primary and secondary pupils need different wording, and pupils with English as an additional language need accessible phrasing and, where possible, audio read-aloud so that reading level does not distort the result.

Staff and family voice

Pupils describe the lesson-level experience, and staff workload and morale shape that experience. If teachers are stretched, relationships and feedback suffer. Asking staff short questions on workload, support and whether they feel able to do their best teaching gives you the other half of the picture. Read the two sets of results together. A drop in pupil scores for one department alongside lower staff scores in the same area points to a different cause from a drop that appears only on the pupil side.

Families add a third perspective. A simple item on whether their child talks positively about school at home tells you something pupils may not say in a school setting. Differences between pupil and parent answers are worth following up, particularly for specific cohorts.

Always include free text

Scores tell you where. Comments tell you why. One open prompt, such as “What one thing would make school better for you?”, is enough. Comments are where leaders find the unclear routine, the corridor problem or the lesson type that pupils find pointless, and they are usually more persuasive to staff than a percentage.

Worked example: comparing Year 7 engagement across a trust

The following is an illustration, not a real trust or real data. Imagine a six-academy trust that runs the same termly survey in the same fortnight each autumn. Five academies score similarly on the Year 7 statement “I feel part of this school”. One scores markedly lower.

A headteacher summary might have said that transition went well everywhere. The common data says otherwise. The trust education director asks the low-scoring academy to look at the result by cohort: SEND, Pupil Premium, EAL and feeder primary. Most breakdowns look close to the academy average. The feeder primary view does not. Pupils who joined from outside the main feeder schools score much lower than those who came with a large group of classmates.

Before reading any further into this, the academy checks group sizes. If the outside-feeder group is only a handful of pupils, the result is a prompt for conversation, not a finding. Suppose here the group is large enough to report safely.

The free-text comments give the likely cause. Pupils from outside the main feeders mention unclear routines, such as where to go at lunch and what happens when you lose a planner, and several say they did not know which adult to ask during the first half term. These are fixable problems.

The academy then compares the survey with its attendance and behaviour data. Suppose attendance for this group is only slightly lower than for the rest of Year 7 and there is no rise in behaviour points. The honest reading is that the survey may be picking up something early, or may be picking up a feeling that will never become a problem. The data cannot yet tell the school which. That uncertainty is a reason to act in proportion and to keep measuring, not a reason to ignore the result.

The action is modest and specific. A named assistant head owns the issue. The academy sets up a buddy system and a named form-tutor check-in for pupils who joined from outside the main feeders, and it publishes a one-page routines guide. The trust agrees to re-survey the same cohort next term using the same question. If the score moves, the change is probably working. If it does not, the team has learned the cause lay elsewhere. For a fuller walkthrough of this kind of trust-wide comparison, see our guide to measuring belonging in a multi-academy trust.

Turning results into action leaders can evidence

A repeating termly cycle keeps measurement connected to decisions. Measure with the same survey in the same weeks. Segment results by cohort and year group. Act by giving each low-scoring cohort an owner, a specific change and a review date. Review at the next survey to see whether the score moved. A cohort without an owner and a date is just a number on a slide.

How results reach staff shapes whether the survey is trusted. Brief staff before sharing, explain what the questions measure and lead with strengths. Then name one or two priorities. When teachers see the survey as a planning tool rather than a judgement of their teaching, response quality and willingness to act both improve.

For trustees, the value is comparability. If every academy uses the same indicators, the board can see where engagement is stronger or weaker without relying on each headteacher’s summary. Report the same measures in the same format each term, and let the board ask about the outliers in both directions, including what the strongest academies do differently. Our piece on trustee questions on inclusion sets out what good evidence looks like for that conversation.

Read results as signals

Set realistic expectations about what the data can carry. Response rates affect how representative results are, so aim for high participation and report it alongside the scores. Small cohorts produce unstable percentages, so check minimum group sizes before reporting any breakdown and respect anonymity thresholds. Treat a result as a signal to investigate, then confirm it with conversations, lesson visits and your existing data. This kind of triangulated record also supports the inclusion evidence Ofsted now looks for.

Perceptions can run this cycle with MIS-linked distribution, cohort breakdowns and automatic analysis of free-text comments by sentiment and topic.

One year group, one survey, one named owner this term

Start small. Choose a single year group, such as Year 7 or Year 10, and run a short pupil survey before half term. Use specific questions on feeling listened to, having a trusted adult, finding lessons worthwhile and feeling part of the school, plus one free-text prompt.

When the results arrive, check group sizes, look at the cohorts, and read the comments. Then agree who will own the lowest-scoring cohort, what they will change, and the date you will re-measure. That single loop will tell you more about pupil engagement in your school than another term of attendance and behaviour reports alone.

Learn more about our services