Statement from Children’s Commissioner Anne Longfield on this year’s GCSE results
It is clear now that the algorithm used by Ofqual to determine this year’s ‘A’ level and GCSE results is irredeemably flawed. It is not in children’s interests for weeks of appeals and arguments among adults to continue. Nor is it in the interest of schools, which should be focusing on their final preparations for reopening safely to all students in September.
Having looked at the evidence, it seems clear to me that GCSE results this year should be based on centre assessment grades – not the algorithm developed by Ofqual.
Pupils will have another two years in school or training to correct any perceived anomalies
For ‘A’ Levels, the grades produced by Ofqual’s algorithm are already out there and have been dissected by all. It is now difficult to put that genie back in the bottle. Ofqual and DfE must now show urgently how they will take further steps to give reassurance to all students that the system is as fair as possible – especially to disadvantaged students. At the very least, students should have their teacher-assessed grades shown alongside their algorithm grades on any certificates. Universities and employers must then consider both, placing more weight on the former if the latter is wildly different.
If what I am calling for results in more students going into further and higher education than the government planned for, then so be it. Ministers will just have to expand the number of places. This is surely better than the risk of throwing a generation’s life chances away, by thrusting students who didn’t get the required grades into the worst jobs market of their lives.
It is notable that other countries in Europe have managed to find better, more creative and fairer ways than the UK of replacing or managing final school examinations during Covid-19. In due course, I hope the Government and Ofqual will consider the injustices that occur when the efforts, talents and dreams of children are considered to be reducible to the output from a statistical model.