In defence of GCSEs and A-levels

Tony Blair was wrong – exams are the best form of assessment we’ve got.

Calls for the reform of our exam system have been gaining momentum, motivated in part by the disruption of Covid. This month, two independent reports on the future of national assessments in England have been published. One, released today (23 August) by the Tony Blair Institute for Global Change, calls for the replacement of GCSEs and A-levels with international baccalaureate-style continuous assessment and assessments of “collaboration, communication, critical thinking and creativity”. The other, by Sam Freedman from the Institute for Government, unpacked the trade-offs in dismantling or reforming assessments – which are high stakes for both schools and students. It argued that, despite their problems, the costs of major reforms to exams are likely to far outweigh any reliable benefits.

Freedman is right about this. Most critics of GCSEs and A-levels can rightly point to their limitations, but few seem to understand the role that exams play in the wider system, or what the likely consequences of making changes to the assessments would be. Exams, to repurpose a Winston Churchill aphorism, are the worst possible form of assessment, except for all the other forms that have been tried.

It is not sensible to embark on fundamental system change at a time when schools are still reeling from pandemic catch-up, and when the latest round of exam changes have barely been implemented. Instead, we should keep tests but make them incrementally better. Fortunately, there is plenty of room for improvement in the current system.

Exam boards in England have a fine tradition, but in many areas of assessment the rest of the world has moved on. Meanwhile, we are stuck with an approach that is used because “we have always done it this way”. Yes, there are some world-leading practices in English assessment, but there are also areas in which tradition should be abandoned. I will give two examples.

First is the reliability of A-level examination grading. For many students who take A-levels, the exact grade they get really matters; a university place can hang on a single grade in a single subject. So they may be surprised to learn that, according to the education regulator Ofqual’s own research, in most subjects the probability of being awarded the correct, “definitive” grade averages out at about 70 per cent – and for some it’s as low as 50 per cent. Many people in the examining world argue this is as good as we can get, and that the trade-offs in improving that aspect would be unacceptable. I tend to disagree. At any rate, we should be having this conversation, and be open to changing the way we have always done things.

Second is what people in the world of assessments call “backwash” – the way high-stakes exams define what gets taught. There is little good research on how to design exams so that people teach and study for them in ways that are most aligned with what we actually want students to learn. Instead, we blame the accountability system for distorting our perfectly good exams out of shape. But in a high-stakes system, exams must be designed so that, as far as possible, teaching to the test is exactly what we want.

It’s easy to say we ought to “scrap exams”. The hard work is in designing an alternative that works better.