This blog on yesterday’s announcement about this summer’s public exams has been written by Dennis Sherwood, an independent consultant. He has previously produced much of HEPI’s past output on A-Levels, including 1 school exam grade in 4 is wrong. Does this matter?
I’ve just been reading about the proposed arrangements for awarding school exam grades, now that schools have been closed and exams cancelled. The key themes are:
- Teachers will be asked to submit their recommended assessments, which will then be ‘processed’ by the exam boards and Ofqual to determine the awards.
- School league tables will not be published.
To my mind, this is realistic and pragmatic – teachers know their students well and are in the best position to make fair and wise judgements, and the knowledge that those judgements will not impact their school’s league table position removes the corresponding pressure to game the system.
How this might work in practice? I don’t know, for few details have been published. But if I were asked to contribute to the design of the ‘process’, one possibility might be for teachers to be asked to submit, on behalf of each student, a recommended grade from U to 9 for GCSE and from U to A* for A level. The exam boards and Ofqual could then check the overall distribution of any school’s recommended grades against corresponding distributions from previous years, and how A level recommendations compare to GCSE and AS level actuals.
Spot checks might then be carried out on the school’s evidence for any outliers, and perhaps for general due diligence too. Personally, I would not be comfortable discriminating between ten GCSE categories, and A level is only marginally less problematic with seven. My preference would be for, say, four – but let me assume here that the ‘process’ requires teachers’ recommendations to reflect the ten GCSE and seven A level grades currently used.
And then I had an intriguing thought. Might teachers do a better job of identifying the right grades than the exam system?
Although I know of no evidence that teachers get this right, there is real, compelling, evidence that the exam system gets it wrong – evidence that comes from the most authoritative source possible, the exam regulator, Ofqual. Here is an extract from an announcement posted to the Ofqual website on 11 August 2019:
more than one grade could well be a legitimate reflection of a student’s performance
This statement was written in response to an article that had appeared in that day’s Sunday Times about the unreliability of A level grades. Ofqual’s statement is unqualified, and so, presumably, applies to all subjects. And its meaning is clear: any grade actually awarded to any student in any subject is just one of an (undeclared) number of grades, each of which is ‘a legitimate reflection of a candidate’s performance’.
In general, this is important but – in this particular instance – even more so. For if the exam system cannot reliably determine whether a particular student merits, say, a grade 5 or a grade 6 for GCSE Geography or an A or an A* for A level History, why might teachers do any better?
Let me answer my own question: because it is, I think, quite possible, if not quite likely, that a teacher is better able to do this than the exam system. After all, teachers have much more information – information accumulated over their careers in general and information about each individual student gathered over, say, two years in particular.
But there are three conditions that must hold.
- First, that teachers behave with honesty and integrity.
- Secondly, that teachers are trusted.
- And, thirdly, that teachers are not pressured into breaching that trust by the so-called ‘unintended consequences’ – the perverse incentives, of performance measures such as league tables.
Indeed, if teachers are trusted, if teachers earn and maintain that trust, if the performance measures encourage the ‘right’ behaviours rather than the ‘wrong’ ones, then the entire examination process could be transformed: exams would probably not be thrown away altogether, but the balance between a teacher’s fair assessment of years of work and what happens in a few sweaty hours in May would be tipped significantly towards the former.
I believe that this would be a good thing, but others will have different views (great – please post a comment). There is certainly a suspicion, and probably hard evidence, that self-assessment results in bias. But right now, there is an opportunity to carry out an experiment that no one would ever have thought possible: to base candidates’ grades primarily on teachers’ assessments for real.
So there’s the challenge. Teachers, step up to the plate, and prove to everyone that, yes, you can do this with honesty and integrity. Demonstrate, as individuals and a profession, that, yes, you can be trusted.
And if that can be seen to happen this year, that sets a precedent – to do this routinely in the future. Over the coming year, while exams are off the table, there is an opportunity to do some other important things too: for example, to determine ways in which an assessment of a single teacher can be sensibly moderated to address the important issue of bias; to consider, seriously, scrapping GCSEs altogether; and – my own bug-bear – to ensure that the grades for all school exams are fully reliable.
Fundamentally, though, it is vital that the teacher assessments to be used over the coming months are, and are seen to be, honest and trustworthy. Yes, it’s all a matter of trust.