How to choose a wellbeing measurement approach for your school

Every school measures wellbeing somehow, even if the instrument is a teacher's gut feel shared in the staffroom. The question is not whether to measure but whether to do it in a way that can be compared over time, across a cohort, and by more than one person. Gut feel is real data, but it retires when the teacher does.

Once a school decides to measure deliberately, it meets a crowded field. It helps to see that almost everything on offer belongs to one of four families.

The four families

Validated survey instruments. Standardised questionnaires, completed by pupils, parents or teachers, with published evidence behind them. The Strengths and Difficulties Questionnaire is the best known. Their strength is credibility and comparability: a score means the same thing in your school as in the research literature. Their limits are cadence and texture. They are designed for screening at intervals, not for noticing that this half term is going wrong, and their categories are fixed whether or not they match what your staff actually see.

Observational profiles. Structured frameworks a trained adult scores after observing a child. Their strength is depth on the children you are already concerned about, and they carry real weight in specialist conversations. Their cost is time and training, which in practice means they are used on a small number of pupils rather than a whole cohort.

Mood check-ins. Pupils record how they feel, often daily, usually through something quick and friendly. Their strength is pupil voice and early warning: a run of grey days is visible long before anything reaches an incident log. Their limit is that self-report from children is noisy, and a check-in alone tells you something is wrong without telling you what.

Framework-based scoring by staff. The school adopts a wellbeing framework, a set of domains and indicators with defined score points, and staff score pupils against it at intervals. The strength is whole-cohort coverage with professional judgement attached: every child gets a considered score, not just the loud ones, and the same framework serves assessment, target-setting and review. The limit is that the framework is only as good as its definitions and the consistency of the people applying them, which is why anchored score points matter. This is the family Phoenix Ember's framework belongs to, with eight domains and defined anchors for each score.

The questions that actually decide it

Sales conversations in this space revolve around features. The decision, in our experience, revolves around five questions:

• Who does the measuring, and when? Anything that requires a hall, a laptop trolley or a training day will happen twice a year at most. Anything a teacher can do in normal time can happen half-termly.
• Does it cover everyone? If the approach only measures children already causing concern, it automates your existing radar rather than improving it. The children you most need to find are the ones nobody has flagged. We say more about them in our plain guide to SEMH.
• Can it show change? A measure you cannot repeat and compare is a snapshot, not a measure. Before and after is the whole game, for the child's sake and for your evidence.
• Does it produce anything? The measurement is a means. If turning scores into a support plan, a report for a parent, or an EHCP contribution is a separate manual job, that job will be skipped in the busy weeks, which are exactly the weeks that matter.
• Will it survive March? The honest test of any approach is not September, when everything is possible, but the low-energy middle of the spring term. Choose the thing your actual staff, at their actual busiest, will still do.

You are allowed to combine them

The families are not rivals. A common and sensible pattern is a whole-cohort scoring framework as the backbone, check-ins for continuous early warning, and a validated instrument or observational profile brought in for the small number of children where a specialist conversation is coming. What does not work is running four systems none of which anyone fully owns. Pick a backbone, make it sustainable, and add to it only when a real question demands it.

Start smaller than you think

Whatever you choose, pilot it with one year group for one term before rolling it out. You will learn more from one real cycle, including the awkward parts, than from any amount of evaluation by committee. And the pilot's baseline data is not throwaway: it is the start of exactly the longitudinal picture you are trying to build.