The grading fiasco has not finished unravelling

25 Aug 2020, 17:15

Even if next year’s exams go ahead without disruption, the game is up for our rigged grading system, writes Dennis Sherwood

Years ago, there was a TV game show called Tell the Truth. Three contestants presented equally implausible stories, one of which happened to be true. The celebrity panellists had to identify the ‘true’ contestant, and the punchline was “Will the real [whoever] please stand up?”.

This summer’s results fiasco has overwhelmed me with déja vu: Will the real grade please stand up?

For many years, we’ve lived with three contestants in the grading game – mocks, UCAS predictions and exam grades – and accepted a consensus that, of these three, exam results are the ‘real’ grade. Mocks don’t count because we all know they are just a practice run done in different ways in different schools, and UCAS predictions are rife with incentives for teachers to bid up.

Unfortunately, the trust given to exam grades has been misplaced. For at least a decade, about 1 grade in 4 has been wrong, some 1.5 million each year. But nobody knew, because the only benchmark of ‘right’ that might have been possible – a re-mark on appeal – was closed down by Ofqual in 2016 in a deliberate and successful attempt to block the discovery of grade errors.

This catastrophe is totally attributable to Ofqual’s vague and confusing ‘Guidance

This year, Ofqual unwittingly gave birth to a new pretender to the real grade throne – the centre assessment grade, or CAG. It was never intended to make a serious claim to the crown, but events have unravelled in unintended ways.

Imagine if Ofqual had not asked for the CAGs, requiring schools to submit only rank orders. This is not so far-fetched: Ofqual’s consultation document published on 15 April made clear that CAGs were in principle superfluous. The algorithm would have churned away and fudged the now notorious ‘small cohort’ problem behind the scenes. On 13 August, A level results would have been announced, as always. Some candidates would have been happy and some disappointed, as always. Some might have been able to appeal on some technicality, but none on the basis of unfairness or to seek a second opinion, as has been the case since 2016. Life would have gone on.

Ofqual’s big problem this year is that they inadvertently created an entire population of second opinions, not from some remote outsider but, to quote Gavin Williamson, from “teachers, who know their students well”. For every algorithmic grade, there was a corresponding CAG. Every candidate in the land could compare them and, if they were different, ask why.

Four plausible yet flawed stories vie for the title: UCAS predictions, mocks, ‘the algorithm’ and CAGS. Chaos ensues. Will the real grade please stand up?

UCAS grades continue to cause confusion – ­‘predicted grade’, ‘teacher’s grade’ and ‘teacher assessment grade’ are bewilderingly similar yet mean such different things. Mocks were stirred into the brew. The algorithm, intended heir to the ‘real grade’ crown, has become roundly discredited as ever more bizarre anomalies have come to light.

But even the hero, the CAG, is being challenged. Different schools determined theirs in different ways, many attempting to comply with ‘no grade inflation’, others submitting what they believed their students deserved regardless, and some likely opting to “just give A*s and 9s and let the machine sort it out” – only to discover to their amazement that all their students have been awarded A*s and 9s.

This catastrophe is totally attributable to Ofqual’s vague and confusing ‘Guidance’. If their algorithm knew the answer in advance, they should have demanded rank orders only. Having asked schools for CAGs, rules regarding grade inflation should have been precise. Even more importantly, they should have asked schools to submit robust evidence of outliers for scrutiny. They did neither, and so the Shakespearian tragedy has evolved: the proclaimed heir toppled, the popular pretender handed the crown.

But the play is not over yet.

Are the heir’s wicked uncles and aunts lurking in the wings, planning a counter-coup?

Will Duke Robert of Halfon lead the Select (Committee) Few to the rescue?

Will the real grade please stand up?

But what if there isn’t a ‘real’ grade after all? What if what we need is some really fresh thinking?

Your thoughts

Leave a Reply

Your email address will not be published. Required fields are marked *

One comment

  1. Huy Duong

    Hi Dennis,

    Regarding you final question, all types of grades are imperfect indicators. Grades by exams have some advantages. First, the imperfection is random. Whether you’re a prince or a pauper, a random person will mark your paper, and they don’t know who you are. There might be more bias in the imperfection of some of the other types of grades. Second, I think exam grades are reasonably accurate for STEM subjects. Third, I guess society might be used to problem of exam grades not being accurate, eg, could that be a reason why a university might have STEM entry requirements consisting of As and A*s (A*AA is not uncommon for engineering at high-ranked universities), but lower grades for arts and humanities. If the profile for inaccuracy across different subjects suddenly changes, I imagine that it might cause difficulties.