DfE forced to update progress 8 scores AGAIN after GCSE statistics error

Another error in the progress 8 scores provided to schools by the government has been identified, this time related to GCSE statistics.

Schools Week understands some pupils have had their GCSE statistics results double-weighted in their progress 8 score instead of maths, leading to an incorrect overall score for their school.

In other cases, statistics results have not been counted in the “open” buckets for progress 8, which again has affected overall progress 8 scores.

The error was identified after schools checked their scores as part of the annual performance tables checking exercise, which allows schools to log on to a government website to see their national data before it is released to the public. The data went live last Wednesday.

However, the DfE says individual pupil level GCSE statistics results are included in the checking exercise data, so the problem appears only to be with the way overall scores have been calculated, rather than missing data.

A DfE spokesperson said: “We are aware that the inclusion of GCSE Statistics results in the provisional performance measures is not as expected.

“This will be corrected in the publication of our provisional performance tables, and schools will be provided with an update in advance of that publication.”

The incident is the second to hit the checking exercise, after it emerged last week that Pearson had sent incomplete BTEC results to the Department for Education, leading to some schools receiving progress 8 scores up to 0.20 points lower than expected.

The DfE has already said the data will be updated to include the BTEC results before it is released in October.

Duncan Baldwin, deputy policy director at school leaders’ union ASCL, told Schools Week that heads “need to be assured that the correct data is used”, given that the “whole purpose” of the data checking exercise is so that early provisional data can be published to help families choose a secondary school.

“There are issues with provisional data anyway, there are often errors, but I don’t recall errors at a huge national scale like this before,” he said.

“It is very important that what is published is correct.”