Because assessment items reference the same curriculum objectives as planning, weak performance points at specific content rather than at a subject, and intervention impact can be measured rather than asserted.
From marks to information
A UK school generates a great deal of assessment data and comparatively little that changes a lesson. Termly data drops, a percentage, a flightpath, a RAG rating. All aggregate, all arriving after the teaching it describes.
The missing element is granularity. “62% in the autumn assessment” is a mark. “Can balance equations, cannot interpret a rate graph” is information, and only the second changes Monday.
Getting from one to the other requires assessment referenced to curriculum objectives rather than to a paper.
What it covers
- Formative and summative assessment referenced to the same curriculum objectives as planning — see National Curriculum.
- Progress tracking against prior attainment, with trajectory rather than snapshot.
- Group disaggregation by pupil premium, SEND, EAL, gender, prior attainment band and any group the school defines.
- Reading and phonics tracking for pupils at risk of falling behind.
- Small-step progress against a pupil’s own baseline where cohort measures say nothing — see SEND.
- Intervention records with pre- and post-assessment on the same reference.
The intervention question
Every school runs interventions. Most can name them. Considerably fewer can say whether a specific intervention worked, because nobody measured its effect separately from everything else happening that year.
The requirement is not complicated: record who was in the intervention, when, how often they actually attended, and what the assessment said before and after on the same curriculum references. Edves does this against the pupil record, which turns “we ran a reading intervention” into a claim with evidence behind it.
This is directly the evidence gap in the achievement evaluation area, and it is what a pupil premium strategy statement is supposed to demonstrate.
Reports that get used
| Report | Question it answers |
|---|---|
| Objective mastery by class | Which specific content has this class not secured? |
| Taught but never assessed | Where is coverage assumed rather than evidenced? |
| Gap trajectory | Is the disadvantage gap narrowing, and in which subjects? |
| Class divergence | Where do two classes in the same subject and year diverge, and on what? |
| Intervention impact | Did this intervention move anything, for whom? |
| Off-trajectory pupils | Who is drifting, and since when? |
| Item quality | Which of our questions do not discriminate and should be retired? |
The class divergence report is uncomfortable and useful. Where two classes taught the same content by different teachers produce consistently different profiles, that is a professional development conversation, and it is visible in data long before it appears in a learning walk.
Assessment when pupils have AI
Work completed outside supervision no longer reliably evidences the pupil. Three responses, in order of durability:
- Assess process, not only product — drafts and checkpoints carry evidence a final submission does not.
- Require brief defence — a short verbal justification is hard to fake and quick to administer.
- Shift weight to frequent low-stakes supervised assessment — ten short in-class checks give a more reliable picture than one substantial homework of uncertain authorship.
Detection tools are not on this list. See AI in education.
A caution about flightpaths
Many UK schools set target grades from prior attainment and track pupils against a projected line. It is administratively convenient and pedagogically questionable: a pupil below their flightpath may be underperforming or may have been given an unrealistic target from a single prior data point.
Edves supports flightpath reporting because schools use it, and also supports curriculum-referenced mastery reporting, which answers a different and usually better question — not “is this pupil where the model says” but “what can this pupil actually do”. Schools should be clear which question they are asking.