Mark Dalton

Ofsted Explorer ·

What’s new in the Ofsted Explorer

Every school and phase, report card areas in context, a second results measure, your official similar schools beside the replica, and the method drawn out.

Cohort size counts properly

A good question from a fellow head prompted this one. A result from a GCSE year group of 60 pupils moves much more from year to year than one from 250, and the Explorer now says so. The table of 50 similar schools has a cohort column, and the contextual model’s “in line” band widens for smaller cohorts and narrows for larger ones: ±2 Attainment 8 points for a median-sized cohort of 180 pupils, scaling with the square root of cohort size. Each school’s cohort and band are stated beside its results. In 2025/26 the contextual gaps of schools with fewer than 60 pupils spread almost twice as widely as those with 240 or more.

The November update

Over the past two days the Ofsted Explorer has grown from a tool for secondary results into something closer to a working reference for every school in England. Most of the changes came from using it in earnest: looking up schools, noticing where a reading was unclear or a chart misled, and putting it right. This is the whole of that work in one place.

Every school, every phase

The Explorer now finds any school with an Ofsted report card, not just secondaries. Primary, special, alternative provision and nursery profiles show the grade for each evaluation area beside schools of the same phase across England, and beside those whose pupils come from similarly deprived areas. Comparisons always use the school’s own phase, and the sections on GCSE results appear only for secondaries.

Every school now has its own page and its own share image, so a link sent to a colleague opens on the right school. Each profile links to the school’s official record on the DfE’s Get Information about Schools and to its Ofsted reports page. New report cards are read from Ofsted’s reports site each morning, and the page says how recent the data is.

Report card areas in context

Each graded area can now be read on its own, chosen from a picker that shows the school’s grade for every area. For each one, the school sits beside four groups: all secondaries in England, schools in the same area deprivation (IDACI) fifth, schools in the same fifth for the share of disadvantaged pupils, and its 50 similar schools.

The disadvantaged fifths are new. They split the 3,298 secondaries with 2025/26 results by the disadvantaged share of their GCSE year group, and they sit alongside IDACI rather than replacing it. The two measures agree closely but not perfectly (a correlation of 0.81), so seeing both is useful. A national page, Report card areas in context, draws the same comparison across every secondary at once.

Results, read with more care

GCSE results can now be read as Attainment 8 or as the share of pupils achieving grade 5 or above in English and maths. Each is read against the national average, against the school’s 50 similar schools and against a contextual model; Attainment 8 is also read against the last published Progress 8, which has no English and maths equivalent.

Progress 8 now uses the DfE’s own published bands, which combine the score with its confidence interval. The contextual model for English and maths has been refitted so that its expected values always stay between 0% and 100%. Before, a handful of highly selective schools were expected to exceed 100%, which made them look as though they had fallen short.

Your official similar schools

Schools can see their official 50 similar schools in the IDSR. Paste those URNs into the Explorer and it now draws the comparison: how many schools the official list shares with the open replica, how the two groups’ results compare, and how closely each list matches the school’s intake on every factor. The list stays on your device; nothing is saved or sent anywhere.

How the comparisons are built

The method section now shows its working. One figure sets out what each factor counts for in the similar-schools model, using the weights the DfE and Ofsted published. Another shows how a school’s 50 are chosen: its 150 nearest secondaries by weighted distance, with the closest 50 picked out. Rebuilt in the browser from public data, they are exactly the 50 the Explorer lists. A third builds the school’s expected result factor by factor, from a school at the England average to the figure the contextual model gives, so it is clear which parts of the intake move the expectation, and by how much.

For meetings and slides

Any school can be downloaded as a 16:9 presentation slide, either as a one-page summary or for a single report card area. Each slide carries its sources, the date of the latest report card data, a caveat where the replica is used, and the line that governs the whole tool: a reading of the data, not a judgement on the school.

Smaller things that matter

The Explorer has been reworked for phones throughout, with text inset consistently from the edges and tables that stack rather than scroll sideways. Information tips now float and stay on screen on every device. The table of 50 similar schools can be sorted by either results measure and can show any report card area.

What it is, and what it is not

The similar-schools groups are an open replica of the DfE and Ofsted model, built from public data. They behave much as the official model does, but they are not the official lists. The contextual model is my own, built for transparency rather than accountability. None of these measures is a judgement on a school, and none should be read as one.

Open the Ofsted Explorer

Sources. DfE, Key stage 4 performance 2025/26 (provisional, 28 October 2026). DfE, Schools, pupils and their characteristics, January 2026. DfE, Special educational needs in England, January 2026. Ofsted, management information as at 31 October 2026, with report cards published since then read each morning to 17 November 2026. DfE and Ofsted, Similar schools model for attainment: technical note (16 July 2026). If you find an error, tell me and I will correct it.