In a normal year the opening of the window for checking school performance data and the release of updated national key stage 2 results is accompanied by the publication of the primary accountability guidance. Evidently this is not a normal year. Schools downloaded their progress data from the checking portal on 7th September; the guidance followed four days later and much of the methodology outlined in the annexes has not been updated since 2023.
This important document helps anyone working with school performance data to understand how the various measures are calculated. It explains which pupils are included in these measures, how pupils working below the standard of the test are handled, the impact of absence from tests and missing test scripts, and what happens if a pupil is out of year group. Perhaps most importantly this year – which sees the first progress measures since 2023 – the document goes deep into the mechanics of how KS1-2 progress is calculated. Trying to make sense of progress data without up-to-date guidance is like trying to navigate mountains without a map.
Since schools have accessed their progress data, issues have emerged. One, relating to the setting of prior attainment bands, is now well known and has been covered in the TES. The DfE have now identified that the threshold for the classification of high prior attainers (HAPs) was set too low – the KS1 average point score (APS) was set at >=8 rather than >8. This has resulted in a large group of middle prior attainers – including those that met expected standards in all subjects at KS1 – being incorrectly placed into the higher band. One school that contacted me stated that two thirds of the 2025/26 year 6 cohort are categorised as high prior attainers; in previous years it was no more than a third.
The DfE state that this will be ‘rectified in revised data and will show correctly in school level data published in December’ (i.e. when the performance tables are published). But provisional ASP and IDSR reports are due to be released before that, and this will likely cause headaches for schools. These reports are shared with governors, SIPs, consultants, LA advisors and inspectors who may ask awkward questions such as “why haven’t more of the high prior attaining pupils achieved greater depth at KS2?” Teachers will then have to explain the error and present a recalculation to counter it.
The DfE explains that the error ‘does not impact pupil and school level progress scores, it only impacts the low, middle and high prior attainment breakdowns for your school’. This has caused confusion because one might assume that if pupils are incorrectly categorised as HAP then their KS2 scores would be compared to the benchmarks that are too high and the resulting progress scores would be negatively inflated. It is important to note that these broad prior attainment bands are used solely for reporting purposes in the IDSR, ASP, and performance tables; they do not form the basis of progress measures, which involves sifting pupils into 20 prior attainment groups.
Pupils being incorrectly classified as high prior attaining is frustrating – and could result in some difficult conversations – but it has no bearing on the progress calculation. The incorrectly classified pupils – specifically those with a KS1 APS of 8 (EXS in all subjects at KS1) – will also be placed into a narrow prior attainment group (PAG) containing pupils with the same KS1 APS. Their KS2 scaled scores will then be compared against the average KS2 scaled score of pupils nationally with the same KS1 APS. Essentially, performance data has rough and fine filters: the rough filter (low/middle/high) is a broad classification for reports; the fine filter (20 PAGs) is used for progress calculations.
An analogy might be describing runners as fast, medium-paced or slow, for the sake of convenience, but for accurate measurement, we compare on the basis of age and gender. The mistake described above is like classifying all 54 year old men (like me!) as ‘fast’ but still correctly comparing their individual race times to the average time of a 50-55 year old man. ‘Fast’ is the broad prior attainment band; ’50-55 year old male’ is one of 20 prior attainment groups.
On the subject of prior attainment groups, below is a comparison of the 2023 and 2026 benchmarks.
At first glance it looks like there have been huge increases in the benchmarks for some of the prior attainment groups (PAG 17 benchmark apparently increasing by nearly 6 points in reading) but the 2023 and 2026 groups do not align and are not entirely comparable. In many cases we need to compare the 2023 benchmark to the 2026 benchmark in the previous row (e.g compare 2023 PAG 17 to 2026 PAG 16). Increases are therefore not as big as they initially seem and tend to be between 1 and 2 points. Increases in benchmark scores were of course expected. The 2026 year 6 cohort had a very different experience to their 2023 counterpart in terms of disruption to their education and when assessments were made pre- and post-Covid. The 2023 Y6 cohort had pre-Covid KS1 assessments and post-Covid KS2 tests; the 2026 Y6 cohort had post-Covid KS1 assessment and KS2 tests that took place 5 years after the pandemic.
We have to assume that the data is correct and attainment has improved, but the recalculation of prior attainment groups – and the odd ranges of the 2026 groups – makes it very hard to compare expected outcomes at the pupil level. There are some other discrepancies, too. 2023 benchmarks are to two decimal places; the 2026 benchmarks are to one, with the exception of PAG 20. And, bizarrely, PAG 19 and 20 overlap. I initially assumed that there was a typo in the PAG 19 range and that it should be ‘>=9.75 to <10.00’, but PAG 20’s benchmarks are lower than PAG 19’s and are duplicates of the 2023 benchmarks. It, therefore, looks like PAG 20 is a remnant of previous guidance and shouldn’t be there.
This combined with guidance that includes other elements of the 2023 methodology – unchanged standard deviations and extreme negative thresholds – undermines our trust in the robustness of the accountability system and causes us to wonder if the issues go beyond the misclassification of middle prior attaining pupils.
We will have to wait until December to find out. In the meantime, many schools will be viewing their data with some suspicion.
Are you interested in exploring the current state of data in schools and finding out what works and what doesn’t? Join us at one of our 2026-27 Data in School Conferences
Leave a Reply