The Secretary of State has confirmed that GCSEs, A-level and AS level exams will not go ahead in England this summer. While the decision feels premature and disappointing, it was surprising – though not as surprising as the statement that “we are going to put our trust in teachers rather than algorithms.”

To quote the Russian proverb famously deployed by Krushchev during the Cuban missile crisis, what we are likely to end up with here is “trust with verification.”  What form would that verification take? For all the attempts at soaring rhetoric at the end of Williamson’s speech, there is still a lot to be thrashed out, and where assessment is concerned, the devil is always in the detail.

Supporters of exams thought we had a sensible solution: later this month, exam boards were to reveal what topics would be examined, and pupils would prepare for those. The argument against public exams has evidently won: for some pupils, there was no decent or fair chance to learn at all. The problem remains that cancelling exams does not automatically increase fairness, for it is still unfair to compare a disadvantaged pupil’s performance in school with that of someone who has not missed a day of learning, be it online or face-to-face.  

And because it is now January and we are in the middle of a UCAS cycle, with most grades predicted and many offers made, we don’t have the luxury of conjuring from scratch a more broad and creative way of assessing pupils’ achievements in school or beyond it. How then are we to assess grades this year, and on the basis of what evidence? 

"Only a minority of pupils will have done mocks by this point."

Involve as many teachers’ and heads’ groups and unions as possible in consultation - and in execution. 

The message I am hearing is key figures haven’t heard a whisper about what the proposals are which allegedly need “fine-tuning.” These are the people who could provide the personnel who could support boards and centres in ensuring that some kind of national standard and framework is being adhered to. 

Don’t make a mockery of mocks. 

A polite reminder to the DfE policy wonk who is obsessed with mocks last August: mocks are part of formative assessment. They are there as a diagnostic tool, to signpost to pupils where they are on the road to achievement. (And all too often to show them how important it is to turn over all the pages and answer the right questions in the right sections, but I digress….) Public exams are summative assessment; they make a judgement on the pupil’s achievement. Last year, pupils who had received a pointy reckoning during mocks were anxious that this was going to damage their position when CAGs were awarded, and we were able to reassure them that teachers were looking at that performance in a context. Only a minority of pupils will have done mocks by this point, and anyone who marked them would have had in mind that they might once again come into play as CAG evidence. So if we are to come up with anything bespoke or specific, it would have to be a portfolio of evidence. But we should approach this with caution.

Don’t be prescriptive about the contents of a portfolio of evidence for moderating purposes, if this route turns out to find favour.

Here what might seem to work for some GCSEs won’t work for most A-levels. Calling on boards and teachers to invent open book or coursework-esque type exercises against yet another rubric to assess for CAG purposes will create a fresh round of work and headaches, and likely accentuate the learning divide that we are attempting to efface. Flexibility is the key here. Tell schools that they must be prepared to justify their decisions to the boards and leave it there. Which leads us on to the crucial point…

Pre-release grades to centres, and/or award grades on a basis of dialogue between centres and boards.

Last April it was mooted that grades would be released in late June. We missed a trick in not suggesting that this should happen on a confidential basis well ahead of students receiving them. Any negotiation could and should happen well ahead of the end of the summer term. Schools would then have notice of where their grades were heading and this would avoid the shock and awe approach of grade release and the consequent FE and HE chaos which we saw this year. 

"Statistics should begin a conversation, not provide all the answers."

For all Gavin Williamson’s reassurance that we won’t be relying on a mutant algorithm to solve the problem posed in 2021 by a mutant virus, we know that the award of grades will remain in many ways a statistical exercise. Where assessment is concerned, the details need to be very well-thought through - as everyone caught out by last year’s small cohort “detail” discovered too late.

That said, statistics should begin a conversation, not provide all the answers. A key lesson from last August has to be about the need to communicate and consult more widely and effectively at every stage of this process. 

I am grateful to Dennis Sherwood for reflecting on these matters with me as I wrote.