EB / News / Management / Flaws in algorithm warning ignored a month ago
Flaws in algorithm warning ignored a month ago
EB News: 20/08/2020 - 09:03
Due to pressure to prevent grade inflation, Ofqual ignored warnings of flaws in the exams algorithm that left thousands of students devastated.
Reports claim that the regulator was warned at least a month ago that its formula for deciding England’s pupils’ A-level and GCSE results was ‘volatile’ and risked producing erratic outcomes. Despite this, and because of longstanding ministerial pressure to prevent grade inflation, that formula was kept in place - prompting first the mass downgrades for a number of students, and then the government u-turn to rescue the situation.
Education Secretary Gavin Williamson has claimed that he only became aware of the scale of the unfairness of the algorithm when the full details were published ‘over the weekend’, although sources have argued that he himself had raised earlier concerns about the model and sought reassurances from Ofqual.
This morning, The Times reported that Jon Coles, chief executive of the United Learning schools trust, warned Williamson last month that Ofqual’s system would unfairly grade pupils, in a letter and a subsequent call. But Williamson declined to act after receiving ‘reassurances’ from the regulator.
Almost half (46 per cent) of teachers across Europe do not think schools are equipping students with the skills needed for an AI future — a concern shared by two in five teachers in the UK (41 per cent).
An additional £20 million of funding has been allocated to support essential capital repairs and maintenance across schools and colleges in Wales this year.
Now in its 17th year, the Education Business Awards continue to recognise and celebrate the outstanding work, dedication, and achievements of schools and academies across the UK.
The measure, added to the Children’s Wellbeing and Schools Bill, delivers on the commitment made in the government’s manifesto to bring multi academy trusts into the inspection system.